back to overview

EU: Results of stakeholder questionnaire on QSAR models for REACH

In recent years the EU has funded research into developing computer-based methods for evaluating the toxicity of chemicals, called QSAR or “Quantitative structure-activity relationship” models. These computerized models potentially make it possible to evaluate large numbers of chemicals (as required to register them under EU REACH legislation) while also reducing the numbers of tests on animals. The ORCHESTRA project1 aims to promote wider understanding, awareness, and appropriate use of QSAR methods. A survey was conducted among potential users (scientists, academics, consultants, regulators, and industry communities) to identify perceptions and needs in relation to the regulatory use of such “in silico” methods. Insight was gathered on these stakeholders’ current sources of information, their range of tools and practice, and the perceived benefits or barriers to use of QSAR methods, as well as policy needs.  

Two online questionnaires were disseminated2 in Europe and beyond (QI, up to 8 questions; QII, 2-4 questions for a busier audience). In all 62 stakeholders responded and despite this modest number some trends or groupings were clearly identified.  

The outputs showed that most of the respondents not only have used QSAR models but were also keen on taking opportunities to apply in silico methods. 28 models in use were identified; the most quoted being the OECD Toolbox, EPISuite, and CAESAR, all of which are free tools. However, results suggest that economic cost is not a barrier to application. Although varying knowledge demands were expressed by the stakeholder segments, overall a considerable need was expressed for “more information or regulatory guidance” on using and applying these methods. This indeed might be the key to reinforcing QSAR model uptake. 

Specifically, scientists and consultants are interested in both the technical and the scientific aspects of QSAR applications, their main concern being information that can help gauge the level of confidence to be placed in a model. Regulator – who are primary actors of the acceptance of QSAR models – seem slightly less concerned by the scientific validity of models, requiring instead a better grasp of which applications are appropriate in the context of REACH. They seek good understanding of software outputs: the results of the model and their meaning. Finally, industry stakeholders want reassurance that the scientific quality of a given tool will be considered acceptable by regulators, and pointers to the best available models.  

As for fostering use, respondents ratified the case study approach, whose value is to consolidate experience with model use. All sectors found that demonstration by industry of successful actual applications will have a large impact on dissemination of in silico methods. “An important factor in confidence building is for both industry and regulatory bodies to simply start using the tools, to see how they work.” Finally, the ORCHESTRA survey revealed agreement on the importance of involvement and networking in order to mutualise expertise and cooperation and to support industry in the application of the methods. A democratic, evidence-based demonstration, with expert quality assurance in the background, emerges as more likely to trigger greater use than would high-profile leadership or trademarking.  

The findings suggest that more detailed attention by developers and consultants to the process requirements of REACH could be valuable. A good model and its suite of tools are not sufficient if the model is not described and if output components are not transparent.  

Under these insights, ORCHESTRA and another EU-supported project, ANTARES (, promoted a website called VEGA (, providing a highly detailed explanation on the analysis of model results for a given substance. 

Claire Mays* and Emilio Benfenati**
*Symlog, Paris, France;
**Istituto di Ricerche Farmacologiche
“Mario Negri”, Milano, Italy

1 A fully detailed report of the questionnaire findings.
2 Both questionnaires can be viewed online at the project website. Questionnaire I and II yielded 33 and 29 responses respectively, 62 in total.

back to overview
Nach oben