Research

Safety culture is a recent concept that describes the common safety behavior and beliefs among an organization. The term Safety Culture was coined by the International Nuclear Safety Advisory Group after a thorough analysis of the Chernobyl disaster of 1986. The definition of this concept is in constant evolution and is described by the American Chemical Society as:

The organization’s collective commitment, by leaders and individuals, to emphasize safety as an overriding priority to competing goals and other considerations to ensure protection of people and the environment.

Safety Culture goes beyond compliance and rules: it concerns people’s habits, prudent behavior and safety consciousness; in other words, it is how an organization behaves when no one is watching. In the last 20 years it has become apparent that many major disasters around the world occurred within a context of a broken Safety Culture.

Academic Safety Culture

Research in the field of academic safety culture began in the mid 2000’s after a series of accidents in laboratories in the United States. The EPFL created its own taskforce in 2016 to understand, enforce and unify the Safety Culture within the School. Our team, is comprised of Chemists, Physicists, Engineer and Life Scientists. The current model is studying the impact of communication, audits, research, teaching and legal texts on the researcher perception of safety. The specificities of the academic environment reveal strong challenges such as:

  • Rapid personnel turnover
  • Cultural diversity
  • Pressure of research (results, costs, schedule)
  • Heterogenous laboratories activities
  • Individual goal-oriented research

An understanding of the safety culture requires a thorough understanding of all its components. Despite these difficulties, most of the major universities around the world share the same problems. The task can therefore be greatly facilitated and the data from our studies supplemented by scientific journals.

Safety climate and surveys

Safety climate is a concept similar to safety culture, however it focuses on observable behaviours. It is mainly used to describe the expressed ideas, the tools and techniques used in general by the organization in order to confirm its compliance to safety. Surveys are the typical tools used to assess this notion. In 2012, an international survey of researchers’ workplace attitudes and practices was launched by few universities; this opened a new interest in the field and stimulated new directions in safety culture studies.

At EPFL, the Safety Correspondent (CoSec) system has significantly helped increase the efficiency and the quality of safety observations and information between the safety coordinators, safety officers, and the laboratory members. This system has the benefit of providing insights into the safety in each laboratory. It was thus logical to conducted a survey among CoSecs. Information on whether laboratory members feel that their colleagues comply to the safety rules of the School and whether their professor supports the laboratory members in taking time for safety related tasks provides the SCC with directions for safety behaviour improvements. Subsets of the CoSec population defined by position within the School and background indicates a difference of safety perception. This tool therefore enables us to create projects, concentrate its efforts and resources, and set priorities to address the issues brought up by the CoSecs.

Chemicals are essential building blocks of life: all living, breathing creatures consists of chemicals and countless chemical reactions take place in cells. Every building we’ve entered is made of chemicals. Every product we’ve ever bought and used is made of chemicals.

Over the last few decades, the number of chemicals added to foods and other products has skyrocketed. We have created all sorts of plastics that are used in innumerable ways. We add preservatives to foods to keep them fresh. We add chemicals to foods to make them look more appealing. We have made food packaging to keep food fresh. We add chemicals to lotions and beauty products to make them feel, look, and smell nice. The list goes on and on of the ways we have invented and used chemicals. We did all of it for good reasons at the time, but now we are learning that many of those chemicals can cause real harm and especially children may be particularly susceptible to the effects of these compounds. They are smaller so their ‘dose’ of any external chemical ends up being higher and they are still developing, so they can be more at risk of harm (their key organs are undergoing substantial changes and maturation that make them more vulnerable). Moreover, they are likely to ingest more toxicants than an adult due to the hand-mouth activity.

Scientists from all around the world are working to establish safe levels of exposure to the most common used chemicals in food, textiles, plastics, cosmetics, pesticides etc. Recently glyphosate, phthalates, bisphenols, perfluoroalkyl chemicals, perchlorate and parabens gained more and more attention regarding their effects on human health.

Our previous work was focused on data gathering on cosmetics found in supermarkets, shops or other retail outlets and hazards assessment of the ingredients. A preliminary data analysis of ingredients of a specific class of cosmetics (shampooing, sun creams) or a specific class of ingredients among different cosmetics (surfactants) highlighted the presence of undesirable substances. Defining the balance between risks and benefits is the challenge.

Nanomaterials are a natural part of the environment and they are produced in numerous processes. Nanomaterials are classified as natural (produced by nature), incidental (unintentionally produced in a process) and engineered (designed and produced by humans).

As the number of engineered nanomaterials (ENM) used in research increases with an incredible speed, and most of the time their implications on the human health is still unknown. Health and safety specialists and regulatory agencies are continuously faced with the challenge of evaluating the risks involved with these materials. Nowadays there is not enough information about their toxicology. Preliminary scientific results indicate that ENM might have a damaging impact on human health, which makes it important to fully manage the involved risks.

The main steps of risk management include: hazard identification, exposure assessment, risk characterization and risk communication, followed by a re-evaluation of the situation. Our research aims at developing new tools for hazard and exposure assessment.

Nanomaterial hazard

What properties make a nanomaterial hazardous? We know that size, surface charge, solubility and other physicochemical properties influence the toxicity of a nanomaterial. Carbon nanotubes are believed to have similar effects as asbestos, because of their high aspect ratio, and certain metal nanoparticles are known to release toxic ions and increase oxidative stress.

Our group has developed a control banding classification system for evaluating the hazard of ENM when toxicological data is not available. We are working in collaboration with researchers in nanotechnology to follow the trends related to new materials and to include classification of these new materials in our method. Our research covers the impact on human health, the environment as well as potential physical hazards.

Occupational exposure

To be able to determine the risk involved with working with nanomaterials it is necessary to evaluate the exposure. We are constantly exposed to natural and incidental nanomaterials, and it is therefore difficult to measure the number of engineered nanomaterials in the air.

Our research is focused on examining methods to measure the nanomaterials released from procedures employed in laboratory work. We are reviewing existing techniques for measuring the occupational exposure to ENM under research conditions.

Additionally, we are developing a test platform that can be used to perform new procedures and measure the nanomaterial emission without risking personal exposure. The aim is to select techniques that are more suitable for monitoring the exposure of researchers to ENM.

Decision making in risk management is a dynamic process, which starts with problem identification and analysis. It is followed by the identification and evaluation of possible solutions and ends up with the selection and the implementation of best alternatives. To ensure the appropriateness of the decision taken results are compared with initial objectives preceding problem identification. If such objectives are not reached the second circle of the decision making process is initiated, otherwise, it terminates with monitoring and securing fulfilment of objectives.

Tractable well-engineered industrial fields with automatized processes and procedures, where human involvement in a hazardous activity is minimized can hardly be compared with the situation in Academia. Constantly changing processes, the ultimate frequent necessity for work adjustments, permanent human involvement and personal turnover require the implementation of different decision making approaches to resolve safety issues.

To develop such a decision support tool, the way risk assessment is conducted, the type of the collected data, its analysis and results representations have to be changed. To achieve reliability and consistency of risk assessment results, all flows of expert judgements and knowledge uncertainties have to be considered and mitigated to the possible minimum. Apart from financial and time constraint, the most important for this type of decision making objective is the safety level. Analysis and determination of the crucial for the successful and safe performance factors along with analysis of contributors to failures lead to the development of necessary toolbox. This toolbox includes not only assessment of risk reduction potential for proposed alternatives for isolated hazard, but also in the context of the experimental setup and human-safety solution interactions.

Will the same safety measure, used by different people, have the same effect? Unfortunately, not, each time there is human-machine or process interaction, the outcome will be always different, so will be the reliability and efficiency of this safety solution. On the other hand, different types of safety solutions in varying degrees are vulnerable to human factors. The knowledge of the extent of such influences on safety solutions will help not only to select the most reliable for the particular context but to address existing negative issues and adapt solutions to increase general reliability.

The final objective of this work is to develop such a decision support tool for academic and research laboratories. To meet this goal we are modifying approach for the risk assessment, using developed by the group LARA software. Parallel to that, we are conducting various qualitative studies in the field of Safety Climate to establish and define the quantitative description of factors influencing human reliability within human-safety tool interaction.

The project aims to develop a decision support tool selecting the best and appropriate for the specific characteristic of academic and research laboratories safety solutions. It implies the development of simple and suitable for various existing laboratory hazards risk assessment tool, which can provide the decision maker with all necessary information, securing its reliability, consistency and relevance. The part of this project includes studies of the human factors and safety climate due to their high influence on the efficiency of proposed solutions.

The great variety of laboratories and fields of research, the high turnover of people, the lack of statistical data related to accidents are only some of the characteristics that make Universities a unique environment, completely different from any industrial setting. Due to these peculiarities, standard risk analysis methods cannot be directly applied at the Universities.

Our team works on the development of new techniques for the risk assessment and analysis specifically tailored for academia/ research environments.

The fruit of this research is Laboratory Assessment and Risk Analysis (LARA) technique. As illustrated in the chart, LARA allows the identification of hazards as well as the assessment and evaluation of risks. LARA also offers the possibility to prioritize risks based on a Laboratory Criticality Index (LCI). LCI is calculated by combining several parameters: severity, worsening factors, exposure to hazard, accidents frequency rate and hazard detectability.

In collaboration with an external partner a software dedicated to risk analysis is also developed and available at: http://riskeval.epfl.ch

Bayesian Network for risk estimate

The main output of “Laboratory assessment and Risk Analysis” (LARA) method is the “Laboratory Criticality Index” (LCI). Used to prioritize risks, it helps the safety officer with the identification of the most appropriate corrective measures and the budget allocation. The calculation of the LCI is currently based on a Bayesian Network that allows us to take into consideration the uncertainties coming from a semi-quantitative estimation method based on linguistic judgments of experts.

Tailored techniques to protect researchers from exposure to hazardous nanomaterials

As the number of engineered nanomaterials (ENM) used in research increases with an incredible speed, health and safety specialists are continuously faced with the challenge of evaluating the risks involved with these materials. Our team is specialized on occupational exposure to ENMs and in technical advancements in protective measures.

Without exhaustive information on the toxicity of the nanomaterials in use, the precautionary principle is applied. The mitigation procedures that arise with this approach can be quite restrictive, and our research focusses on the development of experiments to determine exposure to ENM and new methods for measuring release of nanomaterials.

Our main research topics are:

  • Development of new risk assessment methodologies adapted for work with ENMs.
  • Collaboration with toxicologists to determine the toxicity of new materials.
  • Set up of new experiments to measure release and exposure when working with ENMs.

 

Occupational toxicology

The occupational toxicology is the investigation of toxicity of chemicals found at the work place or in every day life that could impair workers ‘health and well-being. Risk assessment toxicology involves review of the results of descriptive and mechanistic studies, combined with study of potential exposures to make probabilistic estimations of risk in exposed populations. It provides a better understanding of the impact of toxic materials on exposed populations. This research aims to

  • Recognize adverse health effects of environmental and occupational toxicants in individuals and/or populations
  • Obtain a detailed occupational/environmental history
  • Define an evaluation and management plan for exposed individuals and/or population

 

Bayesian Network

The great variety of laboratories and fields of research, the high turnover of people, the lack of statistical data related to accidents are only some of the characteristics that make Universities a unique environment, completely different from any industrial setting. Due to these peculiarities standard risk analysis techniques cannot be applied directly at University.

Our team works on the development of new techniques for the risk estimation specifically tailored for research environments.

In collaboration with an external partner a software dedicated to risk analysis has been created, the so called “Laboratory assessment and Risk Analysis” (LARA).

The main output of this method is the “laboratory criticality index” (LCI). Used to prioritize risks, it helps the safety officer with the identification of the more appropriate corrective measures and with the budget allocation. The calculation of the LCI is currently based on a Bayesian Network that allows us to take into consideration the uncertainties coming from a semi-quantitative estimation method based on linguistic judgments of experts.

 

 

What is the price of safety?

The acceptable safety level depends on many factors. Usually diapason of risk reduction is determined by ALARP principle. However, it varies depending on the initial risk level. As a part of decision-making process proper allocation of resources will also depend on the willingness to spend certain amount of money proportioned with the desirable level of safety. Hence, Academia is more diverse environment with a higher expectation level of safety in comparison with industry, actual safety investments can exceed planned one. The goal of this project is to determine the relation between investments and desired safety level for different units at EPFL. This correlation is important in order to establish threshold of risk tolerance in each specific case, and make a decision-making process for research laboratories easier.