Not all risk are equal - A commentary on 'Premises for Environmental Risk Management'

The environmental risk management approach

The policy document ‘Premises for risk management’ (OmR*) sent by the Netherlands Government to Parliament in 1989 as appendix to the National Environment Plan has been instrumental in moving environmental risk assessment and management into the centre of interest. The ‘environmental risk management approach’ described in the policy document offers numerical limits for possible damage caused by exposure of people and ecosystems to major industrial hazards, to ionising radiation and to substances. With these limits, termed maximally tolerable risk levels, the Government specified what possible damage was considered to be tolerable. In addition to the maximally tolerable risk levels the Government also established at what risk level it considers further reduction of risk was no longer worthwhile, the so-called negligible risk levels. The formulation and practice of this risk policy has stimulated research and discussion about the most effective method of risk management and at the same time has increased awareness of the influence of environmental factors on the health of people and ecosystems.
Between 1989 and 1994 the environmental risk management approach was modified. Emphasis was placed on maximally tolerable risk levels and on the balance between the costs of further risk reduction and its benefits (the so-called ALARA approach). Furthermore, activities that lead to norms for seldom occurring but serious accidents being exceeded, are no longer automatically banned in situations in which other, major interests are involved. Basic policies -as far as human health is concerned, the establishment of norms based on individual mortality and, as far as ecosystems are concerned, the maintenance of species -were not affected by the public debate.

A committee of the Health Council discusses in this advisory report the basis of the risk management approach in Dutch environmental policy. The report is one of a pair; in a follow-up report the committee will elaborate on the process of assessing and decision making as to risks. In the present report the committee also responds to the questions the Minister of Housing, Spatial Planning and the Environment has directed to the Health Council.

Risk and risk measures

The committee takes a view of the concept of risk much broader than that used in the OmR-document. It considers ‘risk’ as the possibility, with a certain degree of probability, of damage to health, environment and goods, in combination with the nature and magnitude of the damage. The causes of risk are action by man, natural events or combinations of these. In this report emphasis is placed on damage to health and damage to plants, animals and ecosystems. The proposed concept of risk is rather broad; possible damage can take all kinds of forms, with respect to type as well as to magnitude, to time of occurrence as well as to duration and to possibilities of recovery.
Simplification of the concept of risk is unavoidable when decisions are made about the tolerability of risks and therefore about associated actions. In the environmental risk management approach of the Netherlands Government, risk is interpreted in a much more restricted manner than it is by the committee. A description in terms of three risk measures is considered sufficient: individual risk, group risk and the collective risk for ecosystems. The first two measures relate to damage to human health, the third to the possibility of damage to species in ecosystems. According to the OmR-document it is possible to compare risks related to major accidents, to exposure to ionising radiation and to substances in the environment by applying these three measures.
The committee feels that this approach excludes consideration of several important characteristics of various risk problems. Effects on human health cannot always be expressed in terms of mortality: mortality is the effect used in the OmR-document in specifying individual risk and group risk. Furthermore, the judgement of the parties involved is determined by more than merely the chance of death and the chance of affecting species. In such a judgement process, voluntariness, confidence, dread, etc. are important in addition to considerations related to the societal benefits of the particular action that causes the risk. These aspects are not sufficiently taken into account by the three risk measures presented in the OmR-document. It follows that risk assessment according to the Dutch environmental risk management approach is in certain instances inadequate, i.e. in those instances it does not provide an adequate basis for decisions on the tolerability of the risk and on risk management measures.

Numerical norms

In the Dutch environmental risk management approach decisions concerning the tolerability of risks are taken based on numerical norms established for the three measures. These maximally tolerable risk levels have an equal value for the probability of being killed due to an accident in a certain installation and for the probability of death from cancer due to exposure to radiation from a well-defined source and from cancer due to a specific carcinogenic substance: 1 in a million per year of exposure. The highest level at which exposure to a toxic substance is without effect is -in terms of policy -considered equal to this value. For the protection of ecosystems from damage by one of these agents it was established as norm that 95 percent of the species must not be affected. The committee is of the opinion that the introduction of equal numerical norms does not guarantee an equal degree of protection, contrary to the policy objectives. This is due to the importance, in a concrete situation, of risk attributes other than the three considered and because the elaboration of the measures proposed in the OmR-document can differ between cases.
It has been taken as a decision rule, that exceeding the norm established on the basis of the maximally tolerable risk level is not allowed. In principle, actions that cause exceeding of the norm value are not permitted. Exceeding of the norm because of earlier actions demands intervention. This is a rigid system that, in a formal sense, allows only very limited possibilities for considering the economic and social benefits of risk-bearing acts. The committee feels that more flexible decision rules and a differentiated system of norms can better do justice to the differences between varying forms of actions, activities and situations. This could lead to an improved set of priorities and to adequate levels of protection.

Group risk

In the OmR-document group risk is the probability of death of a group of ten or more persons within a short period of time due to an accident. This risk measure has been introduced to express societal disruption. The committee is of the opinion that societal disruption after a serious accident is expressed only to a very limited extent by the risk measure group risk, as is also apparent from a study performed at the request of the Department of Housing, Spatial Planning and the Environment. In addition to the number of victims, a.o., houses becoming unusable to live in and land becoming unfit for agriculture are relevant characteristics of societal disruption. The maximally tolerable level associated with group risk is specified in a form that is inversely proportional to the square of the number of victims that were fatally affected: the maximally tolerable probability of an accident with ten times the number of victims should be one hundred-fold smaller. Use of group risk aims at taking into account the possible number of victims; a characteristic that plays an important role when people decide about risks. The manner in which the norm depends on the number of victims is, however, not supported by empirical studies of risk perception. Finally, the committee feels that the delimitation of the area within which the victims are taken into account requires further attention, especially in case of risks association with non-stationary hazards like transport activities.

Risk management

Risk management that aims primarily at prevention of damage or loss involves in the first place a comparison of options, processes or techniques , the choice of the most suitable option, and, the creation of provisions to minimise inevitable damage, particularly in cases of serious accidents. Testing and comparing to numerical norms for, e.g., a concentration of a substance, a radiation dose or an estimated probability of accidents of a certain type and magnitude play mainly a signalling or cautionary role and can provide an answer to the question: is the risk management system functioning properly? The committee feels that satisfying these kinds of norms is not an indication that a sufficient degree of risk reduction has been reached; this would imply a denial of the dynamic character of risk management and wrongly equate tolerating a given risk to accepting a certain damage. The committee further notes that exceeding a norm that has been established for prevention purposes, does not provide information on the possible damage.

Levels of scale

It was envisaged that the Dutch environmental risk management approach would also encompass other forms of environmental risks. The OmR-document mentions exposure to sound, odour and genetically modified organisms and also points to global climate change. This kind of objective requires that the stratification of the risk problems with respect to space and time, but also with respect to complexity should be better appreciated. Global climate change can be important topics for risk analyses, but these analyses require a totally different set of attributes, measures related to these attributes and criteria to deal with the great uncertainties involved - than does the assessment of radiation levels in the street running along the x-ray room of a hospital.

Estimating risk

Risk assessment requires risk estimates. Risk estimation has qualitative and quantitative characteristics and involves models and expert judgement. The committee is concerned about the intention of the Government to impose the use of particular models. This could be acceptable for routine decisions, but there is the danger that for more complicated cases reality is adapted to model rather than vice versa. This is all the more a cause for concern because model uncertainty dominates the uncertainty in a risk estimate and rarely can be expressed in a quantitative form. The committee does support the desire by the Government to distinguish between ‘good’ and ‘bad’ models and to follow, as much as possible, a standardised approach when deriving model parameters. This is why the committee feels that each risk estimate should be accompanied by a discussion justifying the choice of model. In this connection, the committee notes in the Dutch environmental risk management approach, the lack of discussion about how to deal with uncertainties in risk estimates, in particular when numerical norms are derived and tested.

Developments elsewhere

In appendices to its report the committee discusses the risk management approaches in the United States, the United Kingdom and Norway. The committee concludes that much can be learned regarding the further development of the risk management policy, from the views held elsewhere. This is particularly the case with respect to the integration of risk management awareness in all operations and the attention to the quality of the organisation, that has been specified in the Norwegian regulations for the off shore industry, and that are also being applied in other branches of industry.
In the US methods of ranking risks in a manner that is scientifically and socially acceptable (‘comparative risk analysis’) are being studied. Comparative risk analysis is intended to improve the setting of priorities by the regulatory agencies; in The Netherlands setting priorities is a key question too and was explicitly mentioned as an aim of the environmental risk management approach. In the US comparative risk projects the focus appears to shift from risk ranking to questions on the distribution of risk between socio-economic population groups (questions on ‘environmental equity’ or ‘environmental justice’). Risk ranking as such is not sufficient in determining political priorities.
The ‘tolerability of risk’ approach of the British Health and Safety Executive shows strong similarities to that of the OmR-document. It is interesting that in this approach the norms for risk in terms of the probability of a certain effect only receives an indicative value. Norms for, e.g., exposure parameters derived from risk norms, must be ‘practicable’. This implies that the derivation of a norm that is in agreement with the ‘just tolerable’ risk level can differ from that in line with a ‘broadly acceptable’ level. The British approach is therefore more flexible and can be applied more readily in different situations.

Questions of the minister of the environment

Chapter 6 of the report offers the replies to the questions asked by the Minister of Housing, Spatial Planning and the Environment. The replies refer to the comments made in the other chapters. The emphasis in the replies, and in the report in general, is on cautioning against too many simplifications which could possibly lead to socially controversial or undesirable decisions and thus do not work in the interest of public health and of the environment. It is important that simplifications are being made in an explicit way. The committee will elaborate further on these views in the report that is to follow.

* OmR is the abbreviation of ‘Omgaan met risico’s’, the title in Dutch of the policy document.