Portal de Periódicos da CAPES

From the Editors

2014; Wiley; Volume: 34; Issue: 7 Linguagem: Inglês

10.1111/risa.12274

1539-6924

Tony Cox, Karen Lowrie, Terje Avens, Seth D. Guikema,

Risk Perception and Management

In 2013, the SRA Council approved a new Specialty Group on Foundational Issues in Risk Analysis. The first part of this issue presents a series of six papers on foundational topics in risk analysis; the first four emerged from a workshop held in Saló, on Lake Garda, Italy, in August 2012. The first paper, by Policy Area Editor Terje Aven and Enrico Zio, summarizes the background and motivation for the workshop and its key conclusions. It discusses the needs, obstacles and challenges for creating a renewed, stronger scientific foundation for risk assessment and risk management to support current and future technological challenges. The paper reflects on the concept of “foundations of risk assessment and risk management” and the challenges therein, and points to research issues for the coming years. The remaining papers in the series discuss several foundational challenges (italicized in the following paragraphs). Next, Sven Ove Hansson and Terje Aven ask whether risk analysis is scientific, if “science” is interpreted broadly as the practice that provides us with the most reliable statements that can be made on subjects covered by the community of knowledge disciplines, i.e. on nature, ourselves as human beings, our societies, our physical constructions, and our thought constructions. They conclude that risk analysis is a scientific field of study, when understood as consisting primarily of (i) knowledge about risk-related phenomena, processes, events, etc., and (ii) concepts, theories, frameworks, approaches, principles, methods and models to understand, assess, characterize, communicate and manage risk, in general and for specific applications (the instrumental part). A model for risk analysis is presented in which science is used as a base for decision-making on risks. It relates the five elements of evidence, knowledge base, broad risk evaluation, managerial review and judgment, and decision to the domains Experts and Decision makers, and to the domains Fact-based or Value-based. The third paper, by Elisabeth Paté-Cornell and Tony Cox, surveys “lame excuses” for poor risk management and suggests ways to move past them by adopting principles for improving risk management. It extends the classic components of risk analysis (risk assessment, risk management, and risk communication) to include risk attribution and learning. There are some similarities here to the fifth paper, but Cox and Paté-Cornell focus more on practical issues and problems that inhibit effective risk management. Their main recommendation is that risk management in practice needs to move beyond blame assignment and excuses and built a cooperative approach to understanding and correcting the root causes of accidents and failures, an approach that needs to include both technical and organizational aspects. In the fourth paper, authors Roger Flage, Terje Aven, Piero Baraldi and Enrico Zio consider how best to represent uncertainties about unknown quantities in a risk assessment context. The authors examine five principal different approaches: Subjective probability, lower and upper probabilities, other non-probabilistic representations (degree of belief, degree of possibility etc.), hybrids of probabilistic and non-probabilistic representations, and semi-quantitative approaches. They conclude that we need to see beyond probability to adequately reflect uncertainties in a risk assessment context, and that more research is needed to make alternative approaches practicable. The wide variety of decision-making situations encountered in practice calls for a unified perspective that allows the use of several approaches – both qualitative and quantitative – for representing and characterizing the risk and uncertainties. Both subjective probabilities and imprecision intervals, as well as qualitative characterizations, may be used to present different types of information and knowledge important for the decision maker. The final two papers in the foundational issues series turn to the vexed issue of how best to define risk. Paolo Gardoni and Colleen Murphy propose expanding the classic definition of risk as a combination of probability and the consequence to include a third dimension, the source of the risk, which can powerfully affect our perceptions of and responses to risks in ways that frequency and severity information alone might not. The source of the risk refers to the process by which society creates or allows it. This expanded definition of risk attempts to capture many of the more qualitative factors recognized as important in past research on risk perception, risk communication, and related areas. The authors highlight the importance of this third dimension and provide a means to translate it to an operational risk scale. Finally, Andretta Massimo discusses a number of definitions of risk and offers a definition grounded in systems theory. The paper defines the risk to a target of interest to be given by the probability of an adverse effect on the target caused by a given level of damage. This definition differs substantially from many prevailing ones, and creates an opportunity for wider discussion by challenging received definitions. This series of papers is the first since the recent inauguration of the Specialty Group on Foundational Issues to be devoted entirely to the foundations of the field, and we expect it will not be the last. Fundamental questions remain about how best to define and measure risk in complex situations to make it useful for informing and improving decisions; how to link risk to causation and theory formation in science; how to represent and communicate risk without saying more (or less) than we really know; and how to reconcile the tensions among risk perceptions, risk calculations, and individual and collective responses, from outrage to apathy, engendered by different risky activities. We hope that these and other foundational issues will be vigorously explored and debated in the pages of Risk Analysis for the foreseeable future. Two papers in this issue deal with aspects of nuclear safety. The first, by Atsuyuki Suzuki of the Japan Atomic Energy Agency, asks whether science and technology can prevent serious accidents, especially those with low probabilities and high consequences. A wide-ranging and deep reflection on the Fukushima experience, trans-science, Perrow's theory of normal accidents, and the larger international history of nuclear power plant safety, leads to the identification of needs for Japan to transform its national safety management institutions and organizations to foster better learning and communication with parties inside and outside Japan. The author is convinced that, as demonstrated by other nuclear power plants that were maintained safely despite being struck by the same tsunami as Fukushima, it is technically possible to manage nuclear reactors to produce power while protecting adequately against extreme natural hazards. But doing so successfully requires organizational and socio-technical competencies that are challenging to create and sustain. Martínez-Córcoles et al. build a structural equations model (SEM) of worker compliance with safety procedures, using data from a survey of employees at two nuclear power plants in Spain. While acknowledging the paradox that complying with safety rules and procedures does not necessarily guarantee the safest outcomes, since rules typically do not anticipate and cover all situations optimally, and since they may create conflicts or ambiguity in the roles that workers are expected to play, the authors conclude that a combination of formalized procedures and empowering leadership can clarify worker roles in safety and improve safety compliance. Two papers deal with risk perception and communication. Scheer et al. examine the very different uses made of the terms “risk” and “hazard” by stakeholders having different roles, such as regulation, hazard reduction, or economic interest in hazardous activities. They compare definitions of risk and hazard proposed by various bodies, from the relatively cogent but non-specific “Function of probability and magnitude of different impacts” used by the Intergovernmental Panel on Climate Change to define risk, to the rather mystifying “Effect of uncertainty on objectives” proposed as a definition of risk in ISO 31000:2009. They compare these to the understandings and uses of “hazard” and “risk” expressed by experts from various stakeholder groups, observing that considerations of fairness, voluntariness of exposure, naturalness of risk, personal control over risk, and so forth matter very differently to members of industry, environmental and consumer organizations, and public authorities; and that members of these groups rank their own and each others’ perceptions of the importance of such factors differently. These differences may represent different semantic, conceptual, strategic/instrumental and control roles for risk and hazard terms used by different stakeholder groups. Knuth et al. compare perceptions of involuntary risks, e.g., from earthquakes, floods, or terrorist attacks, across seven European countries (Germany, the Czech Republic, Poland, Sweden, Spain, Turkey, and Italy). They find that perceptions of such risks are increased when they have been experienced, with some evidence that experiencing risks of one type may increase perceived threats from other types. Perceptions of the risks from various sources also differ among countries, and the authors encourage further studies of such differences using additional countries. Two papers focus on estimation of chemical exposures from available data. Georgopoulos et al. present a new “Tiered Exposure Ranking” (TiER) framework for supporting various characterizations of exposures to multiple chemicals on multiple time scales. They illustrate the resulting flexible, multiple-attribute informatics system for the National Children's Study (NCS), focusing on a “tier 1” application emphasizing use of publicly available demographic/socioeconomic, behavioral, and environmental data in relation to preterm birth and low birth weight; and a “tier 2” application estimating indices of inhalation exposure to pollutant mixtures for NCS counties to support risk characterization for these endpoints. Bogen and Sheehan study exposures to benzene from mineral spirits solvent, concluding that dermal exposures account for about a third of total (i.e., dermal plus inhalation) exposure from a parts-washing task. They use Monte Carlo simulation to characterize uncertainty and variability in dermal exposures. This paper is likely to prove valuable to other practitioners concerned with modeling uptake of solvents through the skin. How can transportation networks such as highways or waterways, telecommunications networks, power grids, pipe networks, and other network infrastructures be made more resilient? What investments, given limited resources, would make them more reliable, survivable, and recoverable, and less vulnerable to catastrophic failures even under stresses that cripple connectivity or capacities? Baroud et al. propose a stochastic model of inland waterway networks for transporting commodities, focusing on the Mississippi River Navigation System as a case study. They model vulnerability and recoverability (characterized by the time to recover throughput capacity) as random variables, and propose a stochastic optimization formulation and solution heuristic for approximately optimizing recovery activity sets based on comparison and ranking of recovery time distributions. This work illustrates the ever-increasing importance and value of the intersection of operations research and risk analysis, which has crucial applications in infrastructure risk assessment and risk management. We encourage further work on practical methods for optimizing risk management of real-world network infrastructures.