Contributions to Risk Analysis: Improving the understanding and characterization of risk following an uncertainty-based risk perspective

Abstract

The present thesis aims to contribute to the development of generic risk science knowledge by exploring how risk and related concepts are understood and characterized. The main aim is to provide a stronger and more rigorous scientific platform for these concepts and approaches, using an uncertainty-based risk perspective as the conceptual foundation.

In general, risk perspectives can be seen as ways of understanding and interpreting the concept of risk. Over time, there has been a discernible shift away from a traditional probability-based perspective on risk towards contemporary definitions, in which uncertainty is highlighted as a key component of risk. The latter type of perspective, often referred to as uncertainty-based, is founded on the acknowledgement that there is a need to see beyond probability when defining and describing risk. According to this perspective, the risk concept is seen as the combination of the consequences of an activity, covering events (A) and their effects (C) and the associated uncertainty (U). From this, we are led to the representation of risk as (A,C,U).

A large part of the research related to generic risk analysis constitutes generating new knowledge on how to best conceptualize, assess and describe risk and its components. More specifically, focus is directed towards the concepts, approaches, principles, methods and models that exist for this purpose. The papers in this thesis are intended to serve as contributions to this work by taking a closer look at some of these concepts and approaches. Papers I, II, VI and VIII take a more detailed approach by exploring the meaning and interpretation of some specific concepts and terms used to characterize and describe the various components of risk, including the events A and consequences C (Paper I), the uncertainty U (Paper II), as well as concepts and terms covering a combination of these (Papers VI and VIII).

The remaining papers take a broader perspective and consider current practices and approaches for assessing, characterizing and handling risk in the context of contemporary issues such as COVID-19 (Paper III), vaccines (Paper IV) and climate change (Paper VII). The insights gained from this work served as the foundation for Paper V, which provides some suggestions for how current frameworks and practices for risk assessment and -handling can be improved.

Paper I explores some of the most commonly applied metaphors for describing surprising, extreme events, including dragon-kings, black swans, grey swans and perfect storms. These metaphors are subject to a wide range of definitions and interpretations in both scientific and nonscientific literature. In the paper, we analyze and evaluate a selection of definitions found in the literature; based on the analysis, we present a structure in which we outline key characteristics of the metaphors and their interrelationships. The structure highlights that knowledge and unpredictability constitute key concepts for the understanding and interpretation of the metaphors.

Paper II focuses on plausibility as a concept used to measure or express uncertainty. The concept is often referred to in situations with large uncertainties and has gained increasing momentum in recent years, partly due to its application in influential works such as the reports produced by the Intergovernmental Panel on Climate Change (IPCC). However, it is not clear what plausibility actually expresses in a risk context. By reviewing and analyzing current definitions and practical applications of the concept, we show how plausibility can be linked to judgments on likelihood and knowledge. While avoiding precision is an intrinsic idea in many applications of plausibility, we argue that a prudent use of the concept requires some degree of precision on both likelihood and knowledge dimensions.

Paper III examines how risk was assessed, characterized and conveyed in relation to COVID-19. We review and discuss the practices and approaches in four different countries and point to key areas for improvement. Drawing on insights from contemporary risk science literature, we provide some suggestions for how current approaches and practices for handling the pandemic risk can be enhanced. In particular, we underscore the need for precision in the understanding and conceptualization of risk and related concepts. Furthermore, we argue that a proper assessment and characterization of risk requires that the uncertainty and knowledge aspects are thoroughly addressed.

Paper IV focuses on the understanding, perception and handling of vaccine risks. While vaccines are considered one of the most significant achievements in medical history, skepticism and concerns persist, often fueled by issues related to trust and safety. Using a contemporary framework for risk conceptualization and characterization as a basis, the paper outlines and discusses key issues in relation to vaccines and vaccination risk. The paper shows how the characterization and communication of vaccine risks can be improved by clarifying how risk and its components should be understood and expressed, to provide fair and honest representations of the risks related to vaccines.

Paper V proposes an adjusted approach to risk management and decision-making. The paper emphasizes the need for a more comprehensive approach to the review and monitoring stage of risk management frameworks, focusing on three main criteria: the introduction of new decision alternatives, changes in risk, and changes in context/values. By systematically assessing these criteria, the approach aims to identify the need for new risk assessment iterations. The paper advocates a dynamic perspective on risk assessment and risk management, acknowledging the evolving nature of risk and uncertainty.

Paper VI addresses the use and interpretation of the term “real risk”. It uncovers two primary usages: one involves statements about the actual risk level, and the other classifies a risk as “real”, suggesting the potential classification of other risks as non-real. Focusing on the latter interpretation, the paper provides a thorough analysis of the term based on two main risk perspectives: probability-based and uncertainty-based. The analysis shows that, regardless of which perspective is used, evidence/knowledge constitutes a key concept for the understanding and interpretation of the term “real risk”. Although a suggested interpretation of the term is provided, the general recommendation is to avoid its use due to the lack of precision. Rather, the paper advocates using terminology based on risk science knowledge.

Paper VII provides a risk science perspective on climate change research, focusing on three main topics: the use of probability models, the application of the precautionary principle, and the significance of risk and resilience. The paper highlights some key challenges and issues related to these topics, giving particular weight to how uncertainties are understood, described and handled. The paper demonstrates how contemporary risk science, with its concepts, principles, approaches, methods and models, can provide guidance to the climate change research in relation to how to conceptualize and characterize climate change risk, and how to represent and express the associated uncertainties.

Paper VIII takes a closer look at the concept of unpredictability. Based on an analysis of its current use and interpretation in scientific literature, we suggest three main conditions for understanding and assessing unpredictability. Using a conceptual framework for characterizing and assessing risk as a foundation, the paper links the concept to the dynamics of time and knowledge, from which two main implications emerge. Firstly, the framework clarifies the distinction between prospective and retrospective assessments of unpredictability. Secondly, the framework underscores the importance of interpreting knowledge as "justified beliefs" rather than "justified true beliefs," acknowledging the limitations associated with the knowledge base supporting the risk assessment.