The world health report

Chapter 5


What strategies can reduce risks to health?

WHO defines the health system to include all actions whose primary intent is to improve health (5) and some activities that improve health fall outside this definition. Examples include reductions in poverty, and improvements in housing and education, which may well reduce exposures to some types of risks but are not primarily designed to improve health. This chapter is concerned mainly with interventions that have the primary intent of improving health.

Some interventions, however, are difficult to categorize strictly using this definition. One set that has traditionally fallen within the remit of public health covers improvements to water and sanitation. Many water and sanitation programmes fall outside the health portfolio, and clearly such improvements do have considerable amenity value outside health. However, clean water and improved sanitation are considered in this chapter because their attributable burden of disease is so significant. It must be noted, however, that although they improve health, many of their benefits are not readily incorporated into a cost-effectiveness framework and should be considered when comparing them with other types of health interventions.

A number of strategies have been used to reduce health risks that are seen as modifiable. They can be categorized broadly as interventions that seek to reduce risks in the population as a whole, and those which target individuals within the population. The former include intervention by governments through legislation, tax or financial incentives; engineering solutions such as the introduction of safety belts in motor vehicles or the provision of piped water; and health promotion campaigns targeting the general public. The latter include strategies to change health behaviours of individuals, often through personal interaction with a health provider; and strategies to change the behaviours of health providers, particularly in the way they interact with their clients.

Genetic screening is a valuable tool for some diseases associated with the risk factors described in this report, but individual genes are not susceptible to manipulation at present. Genetic screening is not considered further in this chapter.

Risk reduction and behaviour

Many risk reduction strategies involve a component of behaviour change. Even engineering solutions, such as the provision of piped drinking-water, will not result in health improvements unless people are willing to use the new source. Social scientists argue that behavioural change first requires understanding (6,7).4 A number of individual preferences or characteristics influence how people translate understanding into health behaviours, including how averse to health risks individuals are and how they value possible future health decrements compared with other competing choices in their lives such as wealth and lifestyle. These preferences are influenced by information and the influence of advertising and marketing.

"Perceived risk" is the subjective assessment of personal disease risk, based on an individual's interpretation of epidemiological and other types of data. There may be a difference between risk perception as an individual and cultural concepts of risk acceptability by society. For example, although driving without a seat belt may be deemed so unacceptable by a society that legislation is enacted to enforce it, individuals within that society may perceive the risk to themselves as trivial and choose not to use a seat belt.

When it comes to risks to health, individuals and societies sometimes prefer to enjoy the benefits of an activity now without thinking about possible future health costs. High consumption of certain types of food, for instance, is perceived by some people to give current pleasure despite the risk of harmful health effects -- to which they give less weight because they will occur in the future.

There is considerable variation in the rate at which people value and assess adverse events that might happen in the future. Some research has indicated that smokers "discount the future" more highly than non-smokers -- for example, a given probability of developing lung cancer in 20 years is given less weight by smokers than by non-smokers (9). People who discount the future more highly value a given future health risk less highly than people who discount the future less highly, even if they have the same information. The question of how technically to incorporate this into the analysis is discussed later but the effectiveness of behavioural modification interventions is clearly influenced by variations in how people perceive the future.

A set of additional factors also influences the way people respond to risk-reduction interventions. Even when people have heard and understood the message that insecticide-treated nets prevent mosquito bites, and wish to use them to avoid both the nuisance value of mosquitoes and the risk of malaria, a number of factors may prevent them from doing so (10). These include the availability and affordability of nets in their locale and their sleeping arrangements (in a house, or on the street). These in turn will be affected by many factors including personal, community and health system characteristics.

One determinant is culture and the social support networks available, sometimes called social capital. Health system and provider characteristics, such as the way the health system is financed (for example, through social health insurance or user charges) or organized (for example, through managed care or a publicly funded system), also influence behaviours and, through them, the costs and effectiveness of interventions.

Individual-based versus population approaches to risk reduction

Two broad approaches to reducing risk were defined earlier. The first is to focus the intervention on the people likely to benefit, or benefit most, from it. The second is to seek to reduce risks in the entire population regardless of each individual's level of risk and potential benefits. In some cases, both approaches could be used at the same time. Focusing on high-risk individuals can reduce costs at the population level because an intervention is provided to fewer people, but on the other hand it might also increase the costs of identifying the group of people most likely to benefit.

Focusing on people who are more likely to benefit has a significant impact on the health of a nation only when there are large numbers of them. For example, lowering cholesterol with drugs is effective in reducing overall mortality in a group of people at high risk of death from heart disease; targeting interventions to reduce cholesterol to the needs of these people focuses the interventions on a group of people likely to benefit.

However, only a small percentage of the population is at high risk of death from heart disease at any given time, and only some of them can be identified purely on the basis of their cholesterol levels. Recent evidence suggests that the group most likely to benefit from cholesterol reduction consists of individuals with combinations of risk factors, such as being male, with ischaemic changes, who smoke, are obese, are not physically active and have high blood pressure and high cholesterol (11). Designing interventions for people with a combination of those risk factors might well prove to be more effective than treating people only on the basis of their levels of cholesterol (12). This form of targeted approach will subsequently be called the "absolute risk approach".

The high-risk approach can be viewed as targeting the right-hand tail of the risk factor curves in Figure 5.1(13). The alternative is to try to shift the entire population distribution of risk factors to the left -- like shifting the distribution of blood pressure for London civil servants in the direction of that of Kenyan nomads. This has the potential to improve population health to a much greater extent than a high-risk approach, while at the same time reducing the costs of identifying high-risk people. On the other hand, the costs of providing an intervention to the entire population would, in this case, be higher than providing it only to people in the right-hand tail. Which approach is the most cost-effective in any setting will depend on the prevalence of high-risk people in the population and the costs of identifying them compared with the costs of the available blood pressure reduction strategies.

The role of government and legislation

Some areas of behavioural change are likely to be adopted relatively easily once information becomes available, assuming that the technology is affordable. Other types of behavioural change will benefit from active government intervention, particularly those where people have high rates of time discount or low rates of risk aversion. Government action is required if the full potential to improve population health through the reduction of alcohol and tobacco consumption is to be achieved, partly because of the addictive nature of these substances. Such action could be through changes in the law or financial incentives and disincentives. Road safety is another area where a significant number of people might not choose to drive safely, or use seat belts or motorcycle helmets, but government action can encourage them to do so, thereby preventing injuries to themselves and to other people.

Increasing prices through taxation certainly reduces smoking (14) even if smuggling increases subsequently (15). A particular focus of this chapter is to explore if this type of government action is cost-effective. In some countries there has been debate about whether governments should play this type of role, and information on the costs and impact on population health are important inputs to this debate.