Risk assessment is typically viewed by government agencies and those in industry as the “sound science” approach to decision-making, in which decisions are made on the basis of what can be quantified, without considering what is unknown or cannot be measured (Diebold et al., 2008). The latter are usually lumped into the category of uncertainty, as will be shown further on in this thesis. Although few scientists will admit it, risk assessment and other “sound science” approaches to decision-making are highly reliant on a number of policy and scientific assumptions, which are frequently unscientific and subjective (Stirling, 2008).
There is, however, a proper, if secondary, role of risk assessment in increasing our understanding of the complexities of environmental harm, as will become clear further on in this chapter. But, as practiced traditionally, risk assessment has often stood in the way of protecting human health and the environment. This can be attributed to the following assumptions on which the conventional approach to risk assessment is based:
a. Risk assessment focuses typically on quantifying and analysing problems, rather than solving them. It asks how much pollution is safe or acceptable; which problems are we willing to live with; how should limited resources be directed? While these are valid questions, they bar more positive approaches and deeper questions: how do we prevent harmful exposures, and how do we move forward from here? (Resnik & Portier, 2008).
b. The practice of scientific risk assessment as way of analysing the possible harm of products and technological inventions is widely used in modern societies. Risk is typically defined as the
“magnitude of a possible hazard” multiplied by the “probability that a hazard will occur” (Stirling & Gee, 2002: 521-533). Hence, the basic steps of risk assessment are to identify the possible hazards associated with a given technological invention and then to calculate the magnitude of each hazard and the probability of it occurring. This exercise is performed by scientists with expert knowledge in relevant fields. While scientists conduct various tasks, such as basic research, product development and policy advice, in the context of policy making the role of science in risk assessment is largely based on the assumption that every hazard can be accurately predicted, and that its respective probabilities can be calculated using scientific methods.
c. The prominent position of expert-led risk-based approaches in decision-making rests upon the generally and widely shared image of science as a process that produces verifiable, reproducible and therefore trustworthy and objective facts and theories about the material world – an image rooted in the modern tradition of the European Enlightenment. This tradition considers reductionism as the best way to reveal facts and theories – expressed in both the methodological belief that the best way to pursue an understanding of complex systems is to break them up into their component parts, and in the ontological belief that the system itself is nothing more than the sum of these components. Consequently, the conventional approach in science has been to study isolated sub-systems under controlled conditions, to use this knowledge to generate an understanding of the system’s function and, by extrapolation and synthesis with other reductive investigations, to predict the future behaviour of the overall system. This image of science and scientific knowledge, born at the time of the European Enlightenment, secures the view that scientific advice and risk assessment deserve a privileged position in the decision-making processes. Policymakers manage risk by evaluating the information and advice given by scientists, and weigh the perceived benefits against the risks. Therefore, a defining characteristic of this approach is the assumption of a clear distinction between factual and objective, expert-led risk assessment, and normative and value-based (i.e. subjective) risk assessment (Felt & Wynne, 2007).
d. Risk assessment assumes “assimilative capacity”, that is, that humans and the environment can render a certain amount of pollution harmless. Eliminating risk altogether is not a plausible outcome of risk assessment. Risk assessment is used to manage and reduce risks, not to prevent them. This deters progress towards moving to cleaner production (Raffensperger & Tickner, 1999) As a first critical observation on these assumptions, it should be pointed out that risk assessments are susceptible to uncertainty. While it is generally assumed in the conventional approach to risk assessment that science yields results that cannot be doubted, the fact of the matter is that even scientifically based risk assessments generate results or outcomes that are extremely variable in that
they often do not determine the nature and magnitude of uncertainties associated with such risks. A case in point is the questions of safety regarding genetically modified organisms (GMOs) that has become problematic to the extent to which testing of a GMO to identify allergens often prove to be not effective, depending on the tests that are used. As a result, the quantitative results of risk assessments of GMOs are often highly variable (Bailar & Bailer, 1999), leaving policy and decision- makers with as much uncertainty, if not more, as existed before the scientific testing begun.
It should also be pointed out that conventional risk assessment is based on a number of different assumptions about exposure, dose response, and extrapolation of these from animals to humans. All these have subjective and arbitrary elements. As a result of this as well, the quantitative results of risk assessments based on animal models are also often highly uncertain, if not extremely variable. As such, this was recognised as a problem by the European Union when it embarked on its European benchmark exercise in hazard analysis. For example, eleven European governments formulated terms and brought scientists and engineers together to work on a problem about the accidental release of ammonia. The result of the exercise was eleven different risk estimates, placing the magnitude of the risk in range from 1 in 400 to 1 in one million. The organisers concluded that “at any step of a risk analysis, many assumptions are introduced by the analyst and it must be recognized that the numerical results are strongly dependent on these assumptions” (Contini, 1991: 87).
At the same time, conventional risk assessment leaves out many variables, especially multiple exposures, sensitive populations, or results other than cancer. Conventional risk assessment is typically geared toward setting single chemical standards and is incapable of analysing the rich and variable mixtures of chemicals found in many communities. It does not adequately take into account sensitive populations, such as the poor, elderly and children, or those already suffering from environmentally induced diseases. It also rarely looks at effects other than cancer, although many environmental health problems involve respiratory diseases, birth defects, and nervous system disorders. Risk assessment is furthermore designed to analyse linear response (more exposure leads to more harm), and is stymied if this is not the case (Rom, 2006). For example, emerging evidence about the ability of some synthetic substances to disrupt the hormone system in humans is showing that lower doses, rather than high doses, may lead to these effects(Poongothai et al., 2007). So the notion of a ‘dose makes poison’ may not always be spot on. It is not only large doses that cause disease, but small amounts may also induce serious disease.
It is therefore clear that the assumptions of conventional risk assessment need to be qualified and contextualized from a platform of a thorough and critical understanding of uncertainty, and the
manner in which uncertainty itself exposes science to certain risks – at least the conventional notion of science.