The World Health Report 2002 and range of possible consequences. Deaths are commonly seen as one of the most important consequences. Attempts were also made to reduce any uncertainties in making the scientific estimates(2). An important consequence of this change was that individual people were now seen as being mainly responsibility for handling their own risks to health, since many risks were characterized as behavioural in origin and, therefore, largely under individual control. This in turn led to the lifestyles approach in health promotion. For instance, a great deal of attention was paid to combating coronary heart disease through health promotion aimed at high-risk individuals, such as increasing exercise and lowering dietary cholesterol, while policies for combating cigarette smoking also emphasized the importance of individual The need for stronger government regulatory controls also became more apparent, with two other important developments. First, governments in many industrialized countries saw their role as law enforcers and passed legislation to establish new and powerful publ regulatory agencies, such as the Food and Drug Administration(FDA)in the USA and the Health and Safety Executive(HSE)in the United Kingdom. Second, increased attention was given to deriving minimum acceptable exposure levels and the adoption of many new international safety standards, particularly for environmental and chemical risks. This included, for example, risks associated with air pollutants, vehicle emissions, foods and the QUESTIONING THE SCIENCE IN RISK ASSESSMENT The so-called scientifico quantitative approach to health risk assessment aims to produce the best possible numerical estimates of the chance or probability of adverse health outcomes for use in policy-making. Although high credibility is usually given to this approach, how valid is this assumption? Why is this approach often seen as more valid than the judgements made by the public or social scientists? Although risk assessment appears to follow a scientifically logical sequence, in practice there are considerable difficulties in making"objective "decisions at each step in the calculations. Thus the risk modeller has to adopt a specific definition of risk and needs to introduce into the model a series of more subjective judgements and assumptions(3, 4) lany of these include implicit and subjective values, such as the numerical expression fo risk, weighting the value of life at different ages, the discount rates and choice of adverse ealth outcomes to be included. For instance, scientific judgements may be needed on the effects of different levels of exposure or which outcomes to include, particularly which disease episodes should be counted among the adverse events During the 1980s, scientific predictions were seen to be rational, objective and valid, while public perceptions were believed to be largely subjective, ill-informed and, therefore, less valid. This led to risk control policies that attempted to"correct"and"educate"the public in the more valid scientific notions of risk and risk management. However, this approach was increasingly challenged by public interest and pressure groups, which asked scientists to explain their methods and assumptions.These critical challenges often revealed the high levels of scientific uncertainty that were inherent in many calculations. Such groups then became more confident, enabling them to argue strongly for the validity of their own ents and interpretation of risks.30 The World Health Report 2002 and range of possible consequences. Deaths are commonly seen as one of the most important consequences. Attempts were also made to reduce any uncertainties in making the scientific estimates (2). An important consequence of this change was that individual people were now seen as being mainly responsibility for handling their own risks to health, since many risks were characterized as behavioural in origin and, therefore, largely under individual control. This in turn led to the lifestyles approach in health promotion. For instance, a great deal of attention was paid to combating coronary heart disease through health promotion aimed at high-risk individuals, such as increasing exercise and lowering dietary cholesterol, while policies for combating cigarette smoking also emphasized the importance of individual choice. The need for stronger government regulatory controls also became more apparent, with two other important developments. First, governments in many industrialized countries saw their role as law enforcers and passed legislation to establish new and powerful public regulatory agencies, such as the Food and Drug Administration (FDA) in the USA and the Health and Safety Executive (HSE) in the United Kingdom. Second, increased attention was given to deriving minimum acceptable exposure levels and the adoption of many new international safety standards, particularly for environmental and chemical risks. This included, for example, risks associated with air pollutants, vehicle emissions, foods and the use of agricultural chemicals. QUESTIONING THE SCIENCE IN RISK ASSESSMENT The so-called scientific or quantitative approach to health risk assessment aims to produce the best possible numerical estimates of the chance or probability of adverse health outcomes for use in policy-making. Although high credibility is usually given to this approach, how valid is this assumption? Why is this approach often seen as more valid than the judgements made by the public or social scientists? Although risk assessment appears to follow a scientifically logical sequence, in practice there are considerable difficulties in making “objective” decisions at each step in the calculations. Thus the risk modeller has to adopt a specific definition of risk and needs to introduce into the model a series of more subjective judgements and assumptions (3, 4). Many of these include implicit and subjective values, such as the numerical expression for risk, weighting the value of life at different ages, the discount rates and choice of adverse health outcomes to be included. For instance, scientific judgements may be needed on the effects of different levels of exposure or which outcomes to include, particularly which disease episodes should be counted among the adverse events. During the 1980s, scientific predictions were seen to be rational, objective and valid, while public perceptions were believed to be largely subjective, ill-informed and, therefore, less valid. This led to risk control policies that attempted to “correct” and “educate” the public in the more valid scientific notions of risk and risk management. However, this approach was increasingly challenged by public interest and pressure groups, which asked scientists to explain their methods and assumptions. These critical challenges often revealed the high levels of scientific uncertainty that were inherent in many calculations. Such groups then became more confident, enabling them to argue strongly for the validity of their own assessments and interpretation of risks