The assumption made in this report is that risk factors, risk probabilities and adverse events can be defined and measured. This is a valid starting point for the quantification of the adverse effects of a range of risk factors and for health advocacy. However, as we have seen above, when interpreting the global burden of risks to health and using this to design intervention strategies, wider perspectives are needed. Evaluating these risks must take place within a much broader context.
People's risk perceptions are based on a diverse array of information that they have processed on risk factors (sometimes called hazards) and technologies, as well as on their benefits and contexts. For instance, people receive information and form their values based on their past experience, communications from scientific sources and the media, as well as from family, peers and other familiar groups. This transfer and learning from experience also occurs within the context of a person's society and culture, including references to beliefs and systems of meaning. It is through the organization of all this knowledge, starting in early childhood, that individuals perceive and make sense of their world. In a similar way, perceptions of risks to health are embedded within different economic, social and cultural environments.
Much of the original impetus for research on perceptions came from the pioneering work of Starr (13) in trying to weigh the risks from technologies against their perceived benefits. Empirical studies of individual risk perceptions had their origins mainly in psychological studies conducted in the USA (4,14). A major early discovery was of a set of mental strategies or rules, also called heuristics, that people use to understand risks (15). An early approach to study and map people's understanding of risks was to ask them to estimate the number of deaths for 40 different hazards and to compare these with known statistical estimates (16,17). This showed that people tend to overestimate the number of deaths from rarer and infrequent risks, while underestimating considerably those from common and frequent causes, such as cancers and diabetes. This finding has obvious implications for control strategies that are focused on many common and widely distributed risks to health. In addition, rare but vivid causes are even more overestimated. Familiarity and exposure through the mass media tend to reinforce these perceptions. However, people's rank ordering by the total number of deaths does usually correspond well overall with the rank order of official estimates.
Risk factors have many dimensions, including a variety of benefits, and certainly risk means far more to most people than just the possible number of deaths. Another pioneering research study, which is relevant to the present analysis of global risks to health, used psychometric testing to measure perceptions of 90 different hazards using 18 separate qualitative characteristics (18). Following factor analysis these hazards were scaled depending on their degree of "dread" and their degree of "unknown risk" (see Figure 3.1, which shows 20 risks selected from the original 90). A third factor (not shown in the figure) related to the number of people involved. Figure 3.1 clearly shows that the most highly uncertain risks, such as nuclear power and pesticides, are the most dreaded, while risks associated with many health interventions and clinical procedures have more acceptable values. For instance, antibiotics, anaesthetics, childbirth and surgery are perceived as being much safer. The higher the dread factor levels and the higher the perceived unknown risks, the more people want action to reduce these risks, including through stricter government regulation and legislative controls. It appears that people often do not make a simple trade-off of benefits against perceived risks. Rather, they want stronger controls against many risks.
Risks that are both highly uncertain and highly dreaded are also clearly the most difficult to predict and control. Two very important factors for dread were found to be global catastrophe and risks that involve members of future generations. The advent of global terrorism and the development of genetically modified foods are two recent examples. Less dreaded risks tend to be those that are individual, controllable and easily reduced. The more acceptable risks are those that are known, observable and have immediate effects. In addition, the more equitable the risks, the more likely they are to be generally accepted.
It is useful to consider perceptions of dread and unknown risk in relation to public health interventions for reducing risks. If risk factors are to be controlled, the interventions should be perceived to have low dread and a low risk of adverse events. Higher risks from such interventions will normally only be accepted by individuals in the higher risk groups. However, population-wide interventions to reduce risk typically have to cover all people, even those at low risk. Thus interventions used in public health programmes need to have low dread and known low and acceptable levels of risk, combined with high safety levels. Typically, vaccination and screening programmes fall into this category, particularly as they are usually targeted at whole populations and involve many healthy people who are at low risk of getting ill and dying. The favourable perception of the public to prescribed medicines, for example, has been attributed to the direct benefits of such medicines and to the trust people place in their safety, achieved through research and testing carried out by medical and pharmaceutical professionals.