Optimizing the implementation of safe and sustainable by design to better enable sustainable innovation

Summary

The concept of safe and sustainable by design (SSbD) combines considerations of human safety, environmental safety, and sustainability. The SSbD framework developed by the European Commission’s (EC) Joint Research Centre (JRC) and the Safe(r) and Sustainable Innovation Approach (SSIA) developed by the Organisation for Economic Co-operation and Development (OECD) are two key examples of how SSbD concepts could be implemented in future regulatory frameworks. SSbD assessment can greatly aid the innovation process to ensure that new chemicals/materials, processes, and products are safe and sustainable throughout the life cycle. Unilever is a global consumer goods company that integrates the latest safety and sustainability science and SSbD principles in the innovation process. In this work, we review the JRC SSbD framework and OECD SSIA, taking into consideration the reflections of stakeholders through a review of published and gray literature, to evaluate to what extent they enable SSbD innovation. We make recommendations on how current gaps and challenges can be addressed to enable maximal uptake by all stakeholders across the value chain. The recommendations are mainly aimed at ensuring that the JRC SSbD framework and OECD SSIA are conceptually robust and addressing the practical elements in the use of the latest scientific methodology, availability of data, tools, guidance, and a standardized approach to interpreting results and decision-making, through short- and longer term efforts.

Graphical abstract

Chemistry; Green chemistry

Introduction

The safe and sustainable by design (SSbD) concept aims to integrate human health, environmental safety, and sustainability considerations into product design processes.^1,2,3 The concept builds on the precautionary principle, aiming for early identification and mitigation of potential risks and impacts of a given chemical/material, process, and product. The SSbD concept borrows principles from related approaches intended to guide material and technology development, such as green chemistry, safe by design, sustainable chemistry, and circular chemistry.^1,4 Efforts to operationalize the SSbD concept have resulted in the development of several technical frameworks or approaches with anticipated application during the innovation process in industrial settings. Two principal proposals for implementation are the SSbD framework developed by the European Commission’s (EC) Joint Research Centre (JRC) and Safe(r) and Sustainable Innovation Approach (SSIA) developed by the Organisation for Economic Co-operation and Development (OECD).^1,5,6,7 The JRC framework and the OECD approach have been widely discussed in the context of the innovation process through case studies and consultation workshops and appear to have the greatest potential for implementation as regulatory tools in the European Union (EU).

The JRC’s SSbD framework focuses on material and chemical innovation together with associated processes and product uses and is envisaged as a means to implement the EU Green Deal (EGD) Chemical Strategy for Sustainability (CSS).^1,5 The OECD SSIA targets innovation of nanomaterials, advanced materials, and nano-enabled products. SSIA is a combination of SSbD and regulatory preparedness (RP), developed to aid regulators in more effectively anticipating and adapting governance to match the pace of knowledge generation and innovation in such materials and products, facilitated by a trusted environment.^6,7 The OECD SSIA efforts are distributed within and outside the EU. Both implementation proposals of the SSbD concept are intended to inform sustainable innovation by identifying strategic points within the innovation process for embedding safety and sustainability considerations in the context of product performance. We anticipate that the SSbD frameworks or approaches currently being developed will be instrumental in the development and application of next-generation chemicals, materials, and products, enabling more sustainable societies. However, this depends on their successful implementation and widespread adoption. Comprehensive industry engagement in further elaboration and piloting of SSbD implementation proposals, as well as tool development and testing, is needed. Frameworks and guidance must be apprised by modern scientific developments and rooted in practicality. Due to the large number of ongoing efforts in the scientific community to operationalize the SSbD concept, it is especially relevant and timely to reflect on the related recent developments to ensure pragmatic translation of theory into practice and adequate bridging of current and future regulatory requirements.

For over 20 years, Unilever has operated on the fundamental principle that all its product and process innovations must be safe by design. This principle has underpinned its product safety governance, focus on developing new scientific capabilities, and investment in building leading-edge industry expertise. Unilever’s innovation processes aim to integrate safety and sustainability in line with the SSbD design principles. As a global consumer goods company and predominantly a user of chemical ingredients and packaging materials, our perspectives and experiences are centered on product-level considerations and assessments. Our experts in environmental sustainability develop and apply environmental sustainability science approaches, using tools such as life cycle assessment (LCA), to estimate environmental impacts of technologies and products, while our expertise in applying modern, scientific methods and approaches, i.e., non-animal new approach methodologies (NAMs) and next-generation risk assessments (NGRAs), is deployed to evaluate consumer, occupational, and environmental safety.^{8,9,10,11,12,13,14}

In this article, we combine our internal company experience of the innovation process,^15,16 non-animal safety and environmental sustainability assessments, together with insights from a wide-ranging review of the gray and academic literature, offering a consumer goods company perspective on the future of SSbD. We evaluate the key similarities and differences in the two main technical implementation proposals identified earlier, establishing conceptual and/or implementation challenges that may limit successful uptake and impact. We then identify ongoing activities and solutions that may address these gaps and outline recommendations for further work and improvement. In doing so, we aim to contribute to the ongoing development of these innovative tools and approaches and encourage other stakeholders to do the same.

Key reflections on the SSbD concept and current implementation proposals

While there have been many attempts to fully or partly operationalize the SSbD concept,¹⁷ the two implementation proposals illustrated in Figure 1 were found to be the most prominent in both the scientific and gray literature. Although the innovation process is defined in broadly similar ways in both, the various assessment steps and information to be considered differ (Figure 1). These two implementation proposals were found to be most comprehensive in combining safety and sustainability aspects, relative to all those reviewed by Caldeira et al.,¹⁷ in terms of their coverage of the sustainability dimension alone. They are also explicitly geared toward aiding regulatory compliance as well as future regulatory preparedness. The key characteristics of each of them have been outlined in Table 1. Through our literature review, we identified several elements across the two proposals that constitute key conceptual and implementation challenges in the operationalization of the SSbD concept. These are explored in the following sections. While the literature has highlighted a need for harmonized definitions,¹⁸ for the purpose of this article, we align with the terminology specified in the JRC framework and methodological guidance with regard to chemicals/materials, processes, and products.^1,5

Figure 1.

Key assessment elements of the JRC and OECD proposals for applying SSbD concept to the innovation process

Stages 1–5 as per Robert Cooper’s Stage-Gate process for innovation. SSbD, safe and sustainable by design; SSIA, Safe(r) and Sustainable Innovation Approach.

Table 1.

Key characteristics of the JRC and OECD SSbD implementation proposals

	JRC SSbD framework	OECD SSIA
Scope	Chemicals and materials in the EU	Nanomaterials, NEPs, advanced materials, current efforts concentrated in the EU
Guidance on applying to the innovation process	Through scoping analysis, system boundaries, iterative and tiered assessment	Through case studies
Regulatory context	To enable EGD’s CSS	To enable regulatory preparedness for advanced materials and NEPs
Absolute safety	Forms the conceptual basis	Not limited by absolute safety
Hazard-based cutoffs	Yes	No
Risk-based considerations	Secondary to hazard-based cutoffs	More risk-based, from early steps or when making trade-off decisions
Absolute sustainability	Yes (ultimate goal)	No
Environmental sustainability	Yes	Yes
Socioeconomic sustainability	Optional step 5 of the assessment	Included within step 2 of the SSbD assessment
Coherence with future regulations	Focused on current regulations	To be enabled through regulatory preparedness
Use of new scientific approaches	None or limited to early stages of innovation	None or limited to early stages of innovation
Data requirements	Extensive	Extensive
Consideration of trade-offs	Not allowed	No guidance
Use of a scoring system	Yes	No
Guidance on tools for SSbD	Yes	Yes

NEPs, nano-enabled products; EGD, European Green Deal; CSS, Chemical Strategy for Sustainability.

Conceptual basis of SSbD frameworks and approaches

The JRC framework presents an ambition to transition from relative to absolute safety by prioritizing the minimization of hazards before evaluating use and exposure. This ambition is to be achieved through hazard-based cutoffs implemented in step 1 of the assessment framework.¹ The concept of safety is a function of hazard and exposure, which together define the risk posed by the use of a chemical/material, process, or product. This is particularly pertinent to products, as they have a defined use. As such, it is necessary that both elements of safety (hazard and exposure) are considered simultaneously at the outset. The OECD SSIA adopts a more risk-based approach by prioritizing hazard reduction while also considering safe use, either from the early steps of the assessment onward or while interpreting the results and making trade-off decisions. This allows for evaluating the product as a whole, without needing to explore extensively on alternate ingredient choices, if any of the product’s individual ingredients fail the hazard criteria initially. Such an approach is aligned to existing regulations, such as the Classification Labeling and Packaging (CLP) legislation, where test data from a product formulation can replace certain ingredient-based calculations.

Many industrial organizations and academic groups argue that innovation decisions based on hazard assessments, without consideration of the potential to demonstrate safe use, would significantly curtail future chemical and material innovation, possibly limiting consumer, societal, and environmental benefits.^{4,19,20,21,22} A pertinent example is that of enzymes, which are used in various applications such as laundry and cleaning products, food and feed industry, textile processing, etc. Enzymes pose a respiratory sensitization hazard. Yet, their safe use, for example, in laundry detergents continues to be demonstrated by application of efficient exposure control measures, such as stabilizing the enzymes through granulation and minimizing consumer exposure.²³ Enzymes used in laundry detergents offer significant benefits such as low temperature laundry washes, improved washing performance, and reduced use of surfactants. They are key in enabling the green transition of the EGD by contributing positively across several SSbD dimensions.^3,24,25 While the current JRC SSbD framework neither permits nor prohibits substances in general and gives recommendations on how to proceed if the first safety criterion H1 is not met, enzyme-based cleaning products evaluated under the framework would still be identified as candidates for substitution. Without having to substitute or redesign, enzyme-based cleaning products can progress through the JRC framework when essential use is proven. However, the list of elements describing the guiding criteria “necessary for health and safety” and “critical for functioning of society” is non-exhaustive, and hence, the qualifying criterion for essential use of enzymes in cleaning products is unclear.²⁶ Furthermore, if the framework were to become more pivotal in a regulatory context, it would be challenging to advance such enzyme-based formulations through a hazard-focused SSbD assessment using the recommended guidelines, discarding innovations like enzymes that are used safely despite the hazards and that bring significant environmental benefits. In Europe, such a consequence would contrast with current priorities to emphasize innovation and growth.²⁷ Furthermore, since individual substances are generally used in a formulation, there is greater scientific value and balance in focusing on the assessment of products, rather than focusing on the intrinsic hazard of a single ingredient.^28,29 Industry-specific guidance created by the European Chemical Industry Council (Cefic), in support of implementation of the JRC’s framework, also recognizes the importance of considering exposure.³ The Cefic SSbD guidance proposes the consideration of exposure to an ingredient within a formulation or product, and if exposure is effectively controlled, the risk from the adverse effects is significantly reduced. Such risk-based considerations are enabled through a key activity on trade-offs embedded in Cefic’s assessment guidance, spanning across all stages of innovation.³

In comparison to the notion of absolute safety, the concept of absolute sustainability is arguably more pertinent and aligns with the long-established paradigm of strong sustainability.^30,31 However, the lack of practical, fit-for-purpose robust methodologies remains a significant barrier to the implementation of absolute sustainability in SSbD. The planetary boundaries (PBs) concept provides an overarching theory for absolute sustainability, but downscaling it to product design and the innovation process is a challenge. Furthermore, the majority of the PBs have been exceeded at the Earth system level, thereby negating their usefulness for SSbD.^32,33 A more pragmatic and practical approach is to focus on relative comparisons of alternative materials and products with a focus on reduced potential environmental impacts. This is reflected in the evolution of the JRC framework, which introduced absolute sustainability in its first release in 2022 but does not mention it anymore in its 2024 Methodology Guidance, suggesting that this concept is only a long-term goal. By way of comparison, the OECD’s SSIA indicates a desire to move to sustainability assessments that enable innovation to stay within the “safe operating space,”⁷ a term defined in the PB concept.³⁴ The JRC framework describes absolute sustainability as the ultimate goal¹ and introduces the need for environmental sustainability thresholds by product type.³⁵ While adaptation of the European Environment Agency’s concept of threshold levels for sustainability to derive these thresholds has been recommended,³⁶ the methodological steps needed have not been stated. For instance, the introduction of impact thresholds for individual products would require quantified thresholds for all PBs and methodological approaches for assigning shares of the global safe operating space to individual products. These approaches are nascent and have been found to introduce large uncertainty in product environmental sustainability assessments,³⁷ with no imminent consensus on assignment principles.³⁸ Practical implementation and workability of that concept in the short term has been questioned by many in the literature,^39,40,41 as well as in the CEFIC guidance on SSbD assessments.³

Scientific methods and approaches

A proactive consideration of how SSbD can ultimately be implemented using a risk-based approach with adoption of NAMs and NGRAs from start to finish, especially in later stages of innovation, is needed in both frameworks. Significant progress has been made in leveraging these scientific advances to develop NAMs and apply non-animal NGRA approaches for protecting people and our environment from chemical exposures.⁴² Even though both implementation proposals do not limit the use of NAMs throughout the innovation process, in a practical sense, the user of the frameworks will still have to rely on in vivo studies toward the final stages closer to market launch. This is primarily because both implementation proposals align with current regulatory requirements for market launches, and any innovation process involves consideration of comprehensive end-to-end product development needs, from discovery to launch and even to post-launch stages. Therefore, it underscores a broader need for SSbD proposals to reflect the latest thinking on how regulatory frameworks can evolve to maximize the value of NAM data and exposure information to assess more meaningful human health and environmental protection goals. As current SSbD proposals have simply adopted the safety assessment approaches accepted under existing regulations, the use of NAM data for SSbD assessments is largely limited to OECD Test Guidelines related to local exposure human health endpoints. OECD validation principles for evaluating NAMs and defined approaches (OECD 34) are currently being revisited to ensure OECD approaches to technical validation better address more complex toxicological endpoints, while still enabling mutual acceptance of data of regulatory data.^43,44,45 Implementing a fit-for-purpose validation process for NAMs can enable broader regulatory application of NAMs, including within SSbD implementation proposals. This would allow for the tiered use of NAMs and exposure information, integrated using computational approaches even for later stages of the innovation process within SSbD proposals.^46,47,48

When it comes to sustainability, both SSbD proposals acknowledge the three recognized dimensions of environmental, social, and economic sustainability.^1,6,49 Approaches for assessing the environmental impacts of materials, chemicals, processes, and products are considered essential in these proposals, while methods for social and economic assessment are generally presented as advisory. Indeed, economic implications are already integral to business innovation processes, and while some authors are concerned that social aspects may be forgotten,²⁹ many social outcomes are arguably the product of supply chain operations (e.g., forced or child labor and fair wages) as opposed to product design (e.g., affordability and access to hygiene) and should be acknowledged and dealt with as such. Social risks and impacts result from how the value chain is governed, i.e., “soft systems,” which may include international standards, national regulations, and company policies,⁵⁰ as opposed to what the supply chain delivers (i.e., new technologies governed by “hard systems”). Regardless, the social LCA (S-LCA) approach introduced in some of the SSbD frameworks (i.e., the JRC framework) inappropriately attempts to derive the outcome of these (soft) governance considerations via a hard systems (i.e., governed by thermodynamics) assessment approach, designed to define the mass and energy balance across the life cycle.

Evaluation of the environmental dimension of sustainability using the product environmental footprint (PEF) methodology is recommended in the JRC SSbD framework. Although the PEF offers a standardized approach for conducting environmental LCAs, various implementation challenges exist, and the proposed approach to define and implement PEF has been contested.⁵¹ For instance, there are varying levels of methodological robustness for the different impact categories, leading to important differences in the uncertainty of the results. While this would be acceptable in a hotspot identification exercise, it lacks the scientific appropriateness required for comparative assessments, especially in a (pre)regulatory context like SSbD frameworks (e.g., the USEtox method for ecotoxicity and human toxicity, embedded in the PEF, is recognized to have 1–3 orders magnitude of uncertainty on top of an increasing but still limited data coverage^52,53). Data availability also remains a considerable challenge, and the PEF governance of new life cycle inventories is foreseen to be difficult for the development of novel ingredients given the timelines of the innovation process. The OECD’s SSIA framework allows a larger degree of freedom when it comes to the implementation of LCA and does not specify which standard to apply. Beyond the near-term challenge of implementing standardized LCA, the SSbD frameworks will need to accommodate new approaches when they have been evaluated. These may include new or improved life cycle impact approaches (e.g., unlocked by the use of AI^54,55), spatially and temporally resolved LCA,⁵⁶ and prospective assessment (heuristics through to comprehensive ex ante LCA^57,58). Additionally, new and evolving regulations such as the Green Claims Directive (GCD) and environmental rating ecolabels (EREs) should recommend methods that are aligned with SSbD frameworks or approaches, to avoid misalignments and contradictory messages. While harmonized regulation is very much needed in the field of claims and EREs, the methods are still under development, and there is a lack of consensus—this is an opportunity for alignment of new approaches with SSbD recommended methods.

Data and code availability

Implementation of SSbD will require a large amount of high-quality safety and sustainability data. Data availability challenges are widely acknowledged for both dimensions, especially during the early stages of technology development and/or for novel applications of materials, chemicals, or products.^{35,59,60,61,62,63,64} This presents an obvious challenge for SSbD since the ability to influence the direction of innovation is highest when technologies are at an early stage of development. When more information is available in later stages of innovation, the possibilities for influence are more limited. This phenomenon is known as the “dilemma of control” (or the “Collingridge dilemma”),⁶⁵ and both frameworks implicitly acknowledge this dilemma, defining iterative processes of data collection and assessment (Figure 1). Nevertheless, the iterative processes in both frameworks lack mechanisms to maximize the use of existing data or to facilitate seamless data flow across the value chain (i.e., when data are available but bound by confidentiality), thereby addressing data gaps in the innovation process.

While both JRC and OECD proposals for operationalizing SSbD allow the use of non-animal NAM data, these methods are generally not currently adopted by the OECD (see scientific methods and approaches) and are advised for use during the initial stages of innovation. However, at later stages of innovation, users of both frameworks will rely on the use of existing safety data produced for regulatory compliance, such as Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) and CLP. This poses a challenge as the current REACH regulation exempts certain toxicology data if the expected chemical volume is in the lower tonnage bands. Consequently, most chemicals and materials currently in use on the European market lack comprehensive REACH data packages required for a complete understanding of hazard profiles under the two frameworks. Generating new in vivo data is neither a feasible nor an efficient strategy for the innovation process. Moreover, one of the actions under the CSS is to extend the REACH information requirements, including greater use of standard information requirements based on NAMs that go beyond the currently underutilized annex XI.¹ For consumer products, consumer behavior and use of the products provide key insights into the relevant required exposure data. Therefore, a pragmatic approach would be one that is exposure driven and would serve as an impetus to improve our understanding of the uses and exposures of chemicals within the specified boundaries of the assessment. This has recently been demonstrated in a case study, where Wood et al. have shown that relying solely on exposure and bioactivity NAM data can support worker safety assessment for systemic exposures, and it is suitable for regulatory purposes.¹⁴ While such approaches would maximize the use of existing NAM data sources such as the Chembl database from the European Bioinformatics Institute,⁶⁶ the Integrated Chemical Environment database from NICEATM,⁶⁷ Environmental Protection Agency (EPA)’s ToxCast program,⁶⁸ and the Tox21 consortium,⁶⁹ they also highlight the need for similar data sources related to exposure.

For sustainability assessments, specific and up-to-date life cycle inventories (LCIs) are often missing.^70,71,72 The prevalence of data gaps leads to the practice of selecting LCIs of other materials as proxies and/or the adoption of other practices for gap filling within LCIs. This often involves manipulating available LCIs with other sources of information to better represent processes and materials of interest, and this approach to data gap-filling will increasingly be possible to achieve across multiple LCIs in a systemized way, e.g., using digital solutions such as Brightway. Variability in data gap-filling practices presents a challenge in reproducibility and comparability. Furthermore, there are currently inconsistencies in results generated from different LCI databases due to different method choices and a lack of consensus (see scientific methods and approaches). For this reason, care should be taken when using different databases in an individual assessment, and database providers should document all methodological choices with full transparency. Some of these challenges could be partly resolved through recent and upcoming regulatory requirements mandated within the EU, e.g., Ecodesign for Sustainable Products Regulations (ESPR), which aims to improve the circularity, energy performance, and other environmental sustainability aspects of products placed on the EU market. Furthermore, the development of the standardized LCI database developed by the European Commission as part of the Environmental Footprint project aims to address some of these issues. A key regulatory element of the ESPR is the Digital Product Passport (DPP), and where there is alignment in those data or information packages, this could be utilized for the purposes of enabling SSbD or vice versa. Efforts to maximize the synergy in data requirements, avoiding generation of duplicate or divergent information, are necessary. Finally, AI approaches that involve augmenting existing data with synthetic data to enhance dataset diversity and size, may help realize better models by filling data gaps in LCI.^50,51

For safety, there is a growing need for NAM data, which, unlike traditional in vivo data, are characterized by high volumes (sometimes spanning terabytes in a single datasets), high velocity (generated and stored rapidly), and high variety (numerous types of different in vitro experiments).⁷³ Similarly, for sustainability, there is a need for LCI data to become more regionalized and supplier specific, resulting in an increase in already large databases. This increase in the volume of safety and sustainability data will bring further complexities in data governance to ensure accuracy (errors in data input can lead to incorrect decision), consistency (inconsistent data can lead to misinterpretation), completeness (incomplete data can lead to biased results), timeliness (outdated data can lead to misleading outputs), and compliance with data regulations across different geographies, including privacy and ethical standards. Therefore, there is an urgent need for a global chemicals and materials data ecosystem supported by fit-for-purpose data governance and adherence to Findability, Accessibility, Interoperability, and Reusability (FAIR) principles, crucial for good data management and stewardship.^{18,74,75,76,77} Some of these data governance issues are currently being addressed by initiatives such as the EC’s One Substance One Assessment proposal, project PINK funded under the EU’s Horizon Research Program and the Partnership for the Assessment of Risks in Chemicals (PARC) toolbox.^78,79,80 Beyond data governance, enhancing information flow across value chains is becoming increasingly important.^5,81 This necessitates practical solutions for data confidentiality concerns while ensuring shared responsibility for data quality and sharing across the supply chain so that required hazard, exposure, and LCI data from suppliers, manufacturers, distributors, and retailers can be made accessible.^64,82 One approach suggested in the JRC methodological guidance is facilitating legal paths like “license to use” or “letter of access,” granting the right to refer to the data for specific purposes, e.g., SSbD, across the value chain.⁵ Another example is World Business Council for Sustainable Development (WBCSD)’s Partnership for Carbon Transparency (PACT), which enables suppliers to calculate and share product carbon footprints while maintaining confidentiality of raw data. However, PACT is currently focused on data related to commercially available chemicals and is limited to the one impact category. Finally, having the data requirements for SSbD harmonized with regulations for safety and sustainability areas will benefit future data sharing. It would be preferable to avoid determining data and information requirements in isolation, so that there can be consistent and seamless use of data across multiple regulatory purposes.

Identifying and handling trade-offs

Identifying and dealing with trade-offs is a necessity for the successful implementation of SSbD as they are routinely encountered between the different dimensions included. Various approaches to allowing decision-making based on these multidimensional frameworks have been proposed. The simplest is the prioritization of dimensions based on the company’s specific industry and market exposure.⁸³ Another way is the development of methods combining the different dimensions of SSbD together to reach one or more final scores, usually referred to as “scoring.” While the OECD SSIA does not currently give guidance on how to handle trade-offs or scoring of dimensions, the JRC framework appears not to allow them. Indeed, its scoring requires “green lights” or “pass” in all dimensions for the innovation to be deemed SSbD at market launch, achieved through an iterative process of redesign. The WBCSD’s Portfolio Sustainability Assessment (PSA), an approach proposed to employ SSbD principles in product innovation but also for product portfolio management, also essentially prohibits trade-offs, although an elaborate scoring approach is introduced to this end: A/B/C grades are awarded for each indicator but the score of the worst performing indicator is adopted for the product overall score.^1,84 The CEFIC guidance introduces general recommendations about how to handle trade-offs, such as “never trade on safety,” and recommends using a multi-criteria decision analysis (MCDA) method.³ While MCDA has been a popular method to help visualize and solve trade-offs in SSbD assessments (see also Hristozov et al., Dias et al., and Stoycheva et al.^85,86,87), other methods and tools have been tested in the SSbD literature such as the benefit assessment matrix,⁸⁸ the risk analysis and technology assessment,⁸⁹ the Greenness Grid,⁸³ the percentage of positive contributions vs. alternative,⁹⁰ and the “fuzzy Delphi.”⁹¹ Analysis of the consequences of these different approaches in terms of the outcome of SSbD assessments is lacking, although it is noted in the literature that different methods to handle trade-offs will lead to different conclusions, significantly affecting future innovation choices.⁶³

Scoring approaches developed in the context of SSbD frameworks are not generally aligned with other recommended approaches to scoring in (pre)regulations. For instance, the PEF (recommended for use in both the JRC’s SSbD and the EU GCD) introduces a single score that combines all environmental impacts into one aggregated score through normalization and weighting. However, it seems that within the JRC’s SSbD scoring approach, the PEF scoring approach is not adopted; normalization and weighting are not employed, and results are clustered into the three categories of toxicity-related impacts, resource impacts, and other environmental impacts. This might lead to contradictory results from JRC SSbD-aligned assessments, used to inform innovation decisions, and those based on the PEF method, including scoring (also developed by the JRC) and aligned with the EU GCD, which will potentially be used to inform consumer decisions, through claims and ecolabels.

The generalized lack of (good) guidance on handling trade-offs was a key finding of Caldeira et al.¹⁷ in their review of 31 SSbD-inspired approaches. In the real world, it is highly unlikely that a material or chemical will show benefits in every single dimension related to safety and sustainability, and complex trade-offs between all desirable functions are also likely.^92,93 Several authors suggest that trade-offs between safety and sustainability can be considered acceptable in specific situations, for instance, where safety risk management (see conceptual basis of SSbD frameworks and approaches) allows for adoption of materials with improved environmental sustainability.^19,94,95 In addition to trade-offs between safety and sustainability, trade-offs may occur within these dimensions, such as between different environmental sustainability impact categories, between different endpoints in safety,⁹⁶ or between socioeconomic risk categories (if the socioeconomic dimensions of sustainability are assessed).

Tools, guidance, and training for the uptake of SSbD

To facilitate the operationalization of SSbD, it is essential to have access to a comprehensive range of safety and sustainability tools along with appropriate guidance and training for conducting the assessments. For the purpose of this paper, we define “tools” as aids that help evaluate data or information (generated through scientific methods or approaches) for decision-making. These can include any software applications or digital platforms that leverage technology to perform tasks, as well as more qualitative instruments such as questionnaires. These technology-aided tasks include uploading or retrieval of data (see data and code availability), as well as processing and analyzing input data according to established methodologies (see scientific methods and approaches) toward a specific output. Qualitative instruments, which can also be considered tools, include questionnaires, criteria, guardrails, or guiding principles. Both OECD and JRC recognize the critical role of tool availability and accessibility in implementing their SSbD frameworks. OECD’s SSIA identifies 41 tools that support SSbD for nanomaterials, such as, LICARA NanoScan, SUNDS, and GUIDEnano for assessing overall risk/hazard band; ANSES, Stoffenmanager Nano, and NanoSafer CB for evaluating safety of production processes; and NanoRiskCat, ConsExpo nano, and nanoFATE for consumer exposure assessment.⁶ Similarly, JRC’s framework and case study report provide examples of tools such as VEGA, OECD QSAR toolbox, CHESAR, and Vermeer FCM that assist in safety assessment.^1,97 For the sustainability assessment, both the JRC and OECD SSIA establish LCA as the main tool to assess environmental impacts, and the JRC also suggests S-LCA for the social dimension. The guidance and training for the application of tools in both frameworks have mainly been delivered through the publication of limited case studies.^6,97

With respect to tools, the PARC toolbox represents an ongoing initiative to develop a structured inventory of tools that align with various steps and stages of the JRC’s SSbD framework.⁹⁸ The current version of the PARC toolbox includes tools for hazard assessment, such as AMBIT, EPI Suite, and JANUS; exposure and risk assessment, including ART, ChemSTEER, and ConsExpo; and sustainability assessment tools such as LCA using software including Simapro, Brightway2, and openLCA. The main goals of the PARC toolbox include establishing a digital infrastructure for SSbD assessment, creating a structured workflow for tool usage, providing functional connections between them, and aiding interpretation and follow-up actions.⁹⁸ Although the current version of the PARC toolbox includes a wide array of tools that support safety and sustainability assessment, certain gaps still need to be addressed in the future. For instance, there are no tools for the hazard assessment step (step 1) of the JRC framework for later stages of innovation. These stages currently rely on the existence of REACH and CLP/GHS (Globally Harmonized System of Classification and Labelling of Chemicals) data (see data and code availability) but need to eventually incorporate tools that leverage scientific progress in the development of non-animal approaches for safety assessment.³ Recent examples of hazard identification tools that solely use NAM-based data and could fill the gaps in the PARC toolbox include the US EPA Next-Generation Tiered Testing framework⁹⁹ and the SARA-ICE model.¹⁰⁰ The former integrates high-content assays and in silico tools for systemic toxicity hazard characterization, while the latter provides GHS classification predictions for skin sensitization based on NAMs and historical in vivo data. Additional tools to assess sustainability are also needed, especially to unlock the use of more complex LCA methods, such as ex ante/prospective LCA (e.g., tools to unlock assessments of low technology readiness level (TRL) innovations such as the Piccinno upscaling framework)¹⁰¹ or spatially differentiated LCA (e.g., tools such as InVEST). The need for additional tools for sustainability is also acknowledged by the CEFIC guidance, which calls for tools developed for mainstream uptake.

Beyond tool accessibility, relevant practical guidance and training for understanding of and application of SSbD are recognized to be lacking.^{70,102,103,104,105} Practical knowledge of applying the frameworks and tools to diverse chemicals, materials, and products developed through various innovation models, with pragmatism and flexibility, is needed.^{35,59,85,90,92,106,107} Flexible use of different methods or different elements from different frameworks has been explored in the literature, as a way to apply the SSbD concept to suit specific examples.^108,109 To prevent any imprecise use or misuse of flexibility in SSbD frameworks, minimum requirements criteria have also been suggested for use in safety assessments.⁹⁴ Efforts toward developing certain areas of safety assessment, i.e., exposure science, as a strong scientific field have been recognized to be valuable in enabling the safety component of SSbD.⁶⁴ The IRISS Project (international ecosystem for accelerating the transition to Safe-and-Sustainable-by-Design materials, products and processes), funded under the Horizon Europe program, is an example of a targeted effort to address the need for guiding principles, skills, competences, and education needs, across safety and sustainability, by serving as a collaborative platform.¹¹⁰ Apart from these, there has been a proliferation of projects funded by Horizon Europe, which attempt to practically demonstrate the implementation of SSbD, including identifying case-specific bottlenecks and proposed solutions.^111,112

Recommendations

In the aforementioned sections, we have outlined various conceptual and implementation challenges that need to be addressed to optimize SSbD for innovation. We have also assessed the current provisions and ongoing efforts to address these gaps and present our overarching recommendations from an industry perspective in Table 2. The recommendations go beyond the actions currently proposed or undertaken by various groups concerned with operationalizing SSbD and encompass both short-term and long-term ideas, where applicable.

Table 2.

Challenges, proposed or ongoing solutions, and recommendations for successful implementation of SSbD

Challenges	Proposed or ongoing solutions	Further recommendations
Conceptual basis of SSbD implementation proposals	•Proposed by the JRC framework to permit the use of hazardous substances if essential use is proven •Proposed within CEFIC guidance to evaluate trade-offs to facilitate a risk-based assessment	•Foster the application of “safe use” instead of “absolute safety” •Foster the application of relative environmental sustainability assessment in SSbD •Complementary, investment in the development of a comprehensive research agenda on PB-LCA for portfolio/sector-level assessments: to be used to steer wider innovation strategies as opposed to individual material/chemical evaluation
Scientific methods and approaches	•Ongoing efforts to revisit and reassess OECD GD 34 for suitability in assessing NAMs for regulatory application •Ongoing project NAMs4NANO to review NAMs suited for nanoparticles and their qualification •Ongoing improvements of the PEF method by the JRC to tackle implementation issues through expert working groups	•Enable the adoption of the latest safety assessment approaches (e.g., NAMs) for regulatory use by ensuring evaluation of new methods is implemented at pace alongside updates to international guidelines, i.e., OECD •Develop consensus on harmonized LCA approach, including addressing remaining PEF implementation issues •Develop early assessment approaches for sustainability based on heuristics and/or ex ante LCA (comprehensive research agendas will be required to make these a reality for scalable application in SSbD) •Adopt new sustainability methods and approaches when maturity is reached (e.g., new/improved LCIA, spatially resolved LCA, prospective LCA) •Allow flexibility to incorporate new methods and approaches within the framework/approach as they become available, through ongoing dialogue between supply chain actors and regulators
Data availability	•Ongoing efforts to extend the REACH information requirements including more extensive use of standard information requirements based on NAMs •Ongoing efforts to maximize use of existing data and/or provide an integrative data platform through EC’s One Substance One Assessment proposal, project PINK funded under the EU’s Horizon Research Program and Partnership for the Assessment of Risks in Chemicals (PARC) •Development of sustainability data blueprints and standards for standardization of LCI format and structure, to ease of data circulation, such as life cycle initiative’s GLAD project •Ongoing improvements of the PEF data availability by the JRC to tackle data implementation issues through the EF4.0 database development	•Improve understanding of the uses and exposures of chemicals •Maximize the use of existing NAM data sources such as Chembla and ICEb databases in all stages of the innovation process •Develop a global chemicals and materials data ecosystem that can support data availability for SSbD framework/approach •Implement FAIR data principles through future testing guidelines and guidance documents •Enable flow of data and information across the value chain by resolving confidentiality concerns (federated learning) through, for instance, extension of the PACT initiative to full LCIs linked to anonymized suppliers for sustainability and “letter of access” or “license to use” for safety •Develop LCI data gap filling tools and standards for proxy selection •Develop harmonized databases that are based on the same methods and aligned with regulatory requirements •Harmonize data requirements between SSbD and regulations (e.g., DPP)
Identifying and handling trade-offs	•Proposed multiple ways of handling trade-offs, such as multi-criteria decision analysis, benefit assessment matrix, risk analysis and technology assessment, Greenness Grid, percentage of positive contributions vs. alternative method, “fuzzy Delphi” •Developed scoring approaches, like those of the JRC SSbD framework or the WBCSD’s PSA •Ongoing European projects on how to incorporate trade-offs into the further development of the JRC framework	•Acknowledge the existence of and need to handle trade-offs •Develop and test methods to identify and quantify trade-offs and evaluate them in terms of workability and consequences on the final decision •As a follow-up, develop consensus on preferred methods to handle trade-offs in SSbD implementation proposals •Develop consistent scoring approaches to identify and handle trade-offs across the regulatory landscape, allowing for a harmonized flow of information from design through to customer/consumer communication
Tools, guidance, and training for SSbD	•PARC toolbox is aimed at creating a structured inventory of tools for SSbD assessment, aligned with the JRC’s SSbD framework •EU Horizon’s Project IRISS is aimed at connecting and supporting the SSbD community globally •Ongoing projects like SUNSHINE to develop early-stage qualitative tools for safety and sustainability assessmentsc	•Create an exhaustive inventory of tools aligned to different steps and stages of the framework; tools should be user-friendly, undergo an appropriate review process with maintenance procedures established, and include appropriate guidance, supporting documentation, and learning material to facilitate user understanding and adoption •Design novel tools by utilizing modern science based on NAMs and NGRA frameworks •Create a user-friendly, structured workflow that provides functional connections between the tools, aiding interpretation and follow-up actions •Build a competency center to manage the needs for training, collaboration, creating resources, and information/data management •Generate more sector- and technology-specific case studies •Develop a single framework in such a way that it is suited to different technologies and industries •Include SSbD education in the academic curriculum •Develop tools to facilitate mainstream use of advanced LCA methods, such as prospective or spatial assessments, when these methods reach maturity

^ahttps://www.ebi.ac.uk/chembl/.

^bhttps://ice.ntp.niehs.nih.gov/.

^chttps://www.sunshine.greendecision.eu.

Conclusions

SSbD design principles have the potential to greatly facilitate sustainable innovation and enable regulatory use of the latest safety and sustainability science if implemented in a future-fit way. In the context of rapidly evolving regulatory landscapes and expanding, misaligned regulatory information requirements, there is a strong desire in industry to harmonize SSbD approaches globally and do so in the context of a broader modernization agenda. This makes it imminent and imperative to critically appraise the proposed frameworks prior to implementation, considering all actors in the value chain, to ensure any SSbD framework is both scientifically robust and feasible for all stakeholders to adopt.

In this work, building on our experience of SSbD approaches in the innovation of consumer products, we review the two key proposals to operationalize SSbD and make recommendations on how to address the current gaps and challenges, prior to their implementation. We propose that an ideal SSbD framework should be one that is not bound by concepts of absolute safety or sustainability. Instead, it adopts a risk-based approach linked to ingredient or product exposure based on handling and use conditions for safety assessment and embraces relative assessment for consideration of environmental sustainability. It is one that has the provision to prioritize and assess both safety and sustainability using the latest scientific methodologies. A functional framework should guide the user in navigating the challenges related to data availability and reuse of existing data. A practical framework should acknowledge the inevitability of trade-offs and encourage means to tackle appropriate decision-making in the face of such trade-offs. For such an ideal framework to be adopted successfully, relevant tools, guidance, and training need to be familiarized with stakeholders, and the development of additional resources needs to be prioritized. This ideal SSbD framework not only needs to be simple, workable, and user-friendly but also appropriately harmonized to avoid policy conflicts and ensure alignment with global regulatory requirements. Given that the framework still contains gaps requiring resolution, and that the requirements impact resources and costs, regulatory flexibility is necessary for the harmonization of the framework. It is also essential to maintain ongoing dialogue and communication between supply chain participants and regulators to achieve this pragmatism and flexibility. If implemented well, SSbD can accelerate sustainable innovation and close the gap between the latest safety and sustainability science and current regulatory frameworks.

Acknowledgments

The authors would like to thank Henry King, Maria Baltazar, and Julia Fentem for their helpful technical advice and comments during the preparation of this manuscript.

Author contributions

All authors have made a substantial and intellectual contribution to the scientific work and drafting of the paper.

Declaration of interests

The authors declare no competing interests.