12.5 Disease prevention and control of non-communicable diseases
Oxford Textbook of Public Health
Disease prevention and control of non-communicable diseases
Types of prevention
Prevention and care
Reducing risk factors
Social determinants of health
Environmental risk factors
A life-course approach to disease prevention
Non-communicable diseases in developing countries
Changes during the course of life
Burden of chronic diseases
The economy of prevention
In the year 2000 the Executive Board of the World Health Organization (WHO) recommended the 55th World Health Assembly:
to formulate a global strategy for the prevention and control of non-communicable diseases.
to recognize the enormous human suffering caused by cardiovascular diseases, cancer, diabetes, and chronic respiratory diseases, and the threats they pose to the economics of member states.
to notice that these diseases are linked to common risk factors, namely, tobacco use, unhealthy diet, and physical inactivity, and being aware that these risk factors have economic, social, gender, political, behavioural, and environmental determinants.
Although, in principle the WHO knows how to prevent many of the most important non-communicable diseases in industrialized countries, in general it does not know how to implement this knowledge. Many of the proximal determinants of these diseases are known, and as the most important health determinants operate within the domain of lifestyle factors, behavioural changes will be needed (Roemer 1984; Koplan and Livengood 1994; Wynder and Andres 1994). Of course, these changes should be implemented without violating people’s right to choose their own way of living. The ability to set up a preventive programme that is evidence based is limited, as the more distal determinants of lifestyle factors are not known, except that they are related to education, social conditions, peer pressure, role models, etc.
Lifestyle factors are closely related to our roles as consumers and we must realize that many actors are involved. An enormous amount of money is spent on influencing consumer behaviour on the market through advertising. Many people want to influence our lifestyle, yet there are few epidemiologists or public health workers and their financial resources are comparatively sparse. On the other hand, epidemiologists have a message that is of interest to the public and therefore to the media. Thus they have much more influence than their sparse financial resources would suggest. The problem in passing on the information is that when health consequences are addressed they are placed in a distant future—people are asked to give up habits that give them pleasure at present and no any guarantees are given that these changes will actually prevent the diseases in question. The concept of disease causation is used, which apparently is far away from a common-sense concept and therefore has low credibility (J. Olsen 1993).
Rose’s paradox of prevention (Rose 1992) is well documented. Eliminating risk factors for those at highest risk who benefit the most from this intervention will only have limited impact at population levels as the majority of these diseases come from people with moderate risk who will personally benefit very little by avoiding the risk determinants in question. Furthermore, all individual-based health programmes are expected to have limited immediate effect and may increase social inequalities in health as the well educated are probably more likely to adapt to evidence-based disease prevention and health care than non-educated people.
Although many non-communicable diseases are preventable, a large proportion are not, even under optimal circumstances. Furthermore, many of the diseases we may prevent will not be prevented for ever. Prevention has to do with avoiding diseases at a premature stage, not necessarily preventing the diseases from occurring at any stage in life. In addition, the incidence of non-communicable diseases will increase with increasing life expectancy. Non-communicable diseases increase when clean drinking water is provided, poverty has been reduced, and malnutrition is gradually eliminated. Basic health care will lead to good-quality treatment of infections, and implementation of vaccination programmes will increase life expectancy; therefore the lifelong incidence of cancer, and probably of cardiovascular diseases, will increase. All of this is intended and desirable, but unfortunately this epidemiological transition is often followed by undesirable epidemics of chronic diseases. If elimination of environmental hazards is followed by physical inactivity, a high-fat diet, and increased smoking, the incidence of other diseases increases. The shift from communicable to non-communicable diseases in many developing countries is an achievement that cannot and should not be prevented as it is largely driven by forces that prevent premature death. The challenges lie in reducing the avoidable deaths and disabilities related to non-communicable diseases as much as possible with the available resources. What should be avoided is a community suffering acute diseases related to overcrowding and poverty, as well as chronic diseases related to overeating, alcohol abuse, and smoking.
Although wealthy countries spend 90 per cent of world health resources on 10 per cent of world health problems, history clearly shows that expenditure of health resources on treatment alone is not a very efficient way of reducing the number of non-communicable diseases (McKeown 1965; McKeown and Lowe 1974), simply because many of these diseases are not curable. There is no magic treatment that will make arteriosclerosis go away and it is unlikely that any cancer treatment will ever be able to eliminate the excess cancer mortality associated with smoking. Cancer treatment may in time be able to cure a growing number of cancer diseases, but many opportunities for prevention have been lost because we have waited in vain for this to happen.
Making changes in society that will make it easier to exchange unhealthy habits for more healthy ones will most likely have a large effect on life expectancy. Making it easy, safe, and pleasant to use the bicycle as a common means of transportation increases the number of people who take physical exercise every day, reduces air pollution, and saves fuel for more useful purposes. Accepting non-smokers’ right to avoid passive smoke will reduce smoking habits as role models will have less influence when smoking is not performed in public places. Subsidizing healthy food or taxation on alcohol, tobacco, fat, and sugar may also facilitate a change towards a more prudent diet. One could envisage a taxation system where taxes are partly paid according to how much one’s behaviour has negative consequences for the environment and to what extent one expects to need the health-care system to cure self-inflicted health problems.
Non-communicable diseases have only one common property —the diseases are non-transmittable directly from one person to another via a single external agent. However, transmission may occur via genetic factors from parents to offspring, or indirectly via role models, or by manipulating the determinants of the disease. This process does not necessarily operate from people with the disease to the non-diseased, and often the mechanism taks place via more distant determinants of the disease. Non-communicable diseases include arteriosclerosis, psychosocial diseases, low back pain, infertility, congenital malformations, poor visual acuity, hypertension, psoriasis, diabetes, etc. Some of these diseases may be caused by infection, but most are probably not.
It is not possible to cover all options for preventing these diseases in a single chapter. Therefore only certain aspects of prevention and health promotion will be discussed, and some of the most important risk factors will be presented.
Types of prevention
Prevention is often classified as primary, secondary, or tertiary. Primary prevention aims at lowering the occurrence rate of the event, i.e. the incidence rate of the disease. Secondary prevention aims at lowering the occurrence of the later and more severe stages of the disease, often by identifying diseases at a curable stage, as in screening, thus reducing the prevalence of the disease through treatment. Tertiary prevention aims at reducing the social consequences of the disease.
Although smoking cessation may be seen as a primary prevention of lung cancer, it could also be viewed as secondary prevention if lung cancer is seen as a disease that starts with a first-stage transformation in a multistage carcinogenic process. A screening programme would aim at diagnosing the disease at a stage where radical treatment is possible and thereby removing the patient from the pool of prevalent cases. Tertiary prevention would aim at securing work and income as long as possible and then to provide aid to reduce the losses following the disease, or to rehabilitate the patient to make it possible for him or her to work at a reduced physical or mental capacity.
In many cases, the distinction between primary, secondary, and tertiary prevention is not clear, because the disease has no clear onset in time but is a result of a process which may be lifelong (Kuh and Ben-Shlomo 1997). The distinction between disease and non-disease is not as clear as many believe, not even for many cancers, and certainly not for cardiovascular diseases or mental disorders. Without a clear demarcation it is difficult to base the taxonomy of prevention on the onset of certain stages of a disease, which may be unknown. For convenience, we often use the time when the clinician makes the diagnosis as a surrogate measure for the onset of disease.
For example, the periconceptual intake of folic acid to prevent neural tube defects is primary prevention (Berry et al. 1999). Prenatal screening for neural tube defects by ultrasound examination is secondary prevention if a finding leads to an induced abortion, which removes the child from the pool of prevalent cases at birth. High-quality surgical treatment followed by extensive physical and social training is tertiary prevention, aimed at reducing the social consequences for the affected child and the family.
Screening is usually considered to be secondary prevention as it aims at identifying those with diseases at a time when they will benefit from early treatment. Thus those with diseases are identified before reaching the critical point where only palliative treatment is available with the present technology. Screening consists of a programme, not just the application of a screening test, and it must be evaluated as a programme. Even in situations where a well-accepted screening test is available, a health benefit is expected, and the necessary health-care facilities are available, a screening programme must be carefully evaluated before it is implemented. The reason for this is not only that the screening programme takes up resources that could be spent in other areas, but also that a screening activity often has severe side-effects for some of the participants. When a screening test is applied to people without symptoms it aims at identifying the diseased before the critical point is reached. An illustration of the simplest possible outcome after using a screening test on a population without symptoms shows how participants fall into four groups (Table 1).
Table 1 Illustration of the outcome of a screening test
The true positives benefit from the screening if they are diagnosed before the critical point. Those who are diagnosed late may be harmed by the screening in some cases. The true negatives often benefit also as they do not have the disease and are reassured by the testing. The false negatives may have the normal diagnostic routines delayed as the test was negative, and the false positives may have to go through unpleasant and perhaps risky diagnostic routines due to the screening result. Therefore the validity of the screening test is a key parameter for the success of a screening programme. It is usually measured as sensitivity, i.e. the probability of being tested positive given that you have the disease, and as specificity, i.e. the probability of being tested negative given that you do not have the disease. Screening tests often have a sensitivity ranging from 30 to 95 per cent and a specificity ranging from 80 to 95 per cent. When such a test is applied to a population with a low prevalence of the disease to be screened, many false-positive results are obtained, and not all with the disease are identified (Hugod and Fog 1992). Randomized control trials have shown a reduction in cause-specific mortality after screening for breast cancer and colorectal cancer (van Dam et al. 1995; Kronborg et al. 1996; Gyrd-Hansen 1999; Gotzsche and Olsen 2000; Jatoi 2000; De Koning 2000). It is also generally believed that screening for cervical cancer is useful in affluent societies. It is expected that more useful screening programmes will be available in cancer prevention in the future, but it would be dangerous to rely upon screening only in the fight against cancer.
Screening for hypertension and hypercholesterolaemia in preventing cardiovascular diseases seems to have a favourable cost–effect ratio, at least in high-risk groups (Schueler 1994; Littenberg 1995; Hughes 1997; Hedner 1998; Holloway et al. 1999).
The concept of prevention is as difficult as the concept of causation. Causation usually addresses proximal determinants of a disease because the cause must make a difference for at least some of the exposed (Lyon 1967). For example, we believe that tobacco smoking, asbestos, and some other exposures cause lung cancer, and by using counterfactual reasoning we believe that if we remove these causes we eliminate some of the expected cases from occurring. Very few of the proximal determinants are in themselves sufficient or necessary causes. Most of the necessary causes are made necessary by including the cause in the definition of the disease (J. Olsen 1993). However, the component causal model presents a framework for causal thinking that matches actual evidence (Mackie 1974; Rothman 1976).
Murray and Lopez (1999) have estimated that the five most important risk factors worldwide are malnutrition, tobacco, hypertension, poor water, and physical inactivity (in that order). These risk factors account individually for 11.7 per cent to 3.9 per cent of all deaths. Prevention should, if possible, aim at these proximal determinants directly, but in practice we may be more successful if we try to modify the determinants of, say, malnutrition or smoking (McMichael 1999). The first of these risk factors is to a large extent associated with poverty, and the second is to some extent associated with ignorance and powerful disease-provoking advertisement.
Although it is believed that specific subtypes of human papilloma virus are necessary causes of cervical cancer, dealing more distal risk factors, such as reducing the number of sexual partners or encouraging condom use, may be the only way of preventing cervical cancers while we wait for an efficient vaccine to be available for the public. Whether we should move upstream or downstream in the search for public health determinants remains an important topic for discussion (Olsen et al. 1999).
Health promotion includes activities that aim at improving health rather than preventing specific diseases. A prudent diet, physical exercise, better social networks, and a stimulating work environment will probably improve well being and lower the risk of several diseases.
In the past in developed countries, and now in many developing countries, communicable diseases were reduced by providing safe drinking water, avoiding crowding by reducing family size, and improving housing conditions and nutrition. Vaccination programmes have sustained this effect in developed as well as in developing countries (Holland 1995).
Preventing non-communicable diseases will, with our present technology, rest upon encouraging healthy habits such as non-smoking, more physical exercise, a better diet low in fat and rich in fruit and vegetables, and better stress control by improving social networks (WHO Europe 1999). However, poverty-related problems such as homelessness, drug and alcohol abuse, and physical violence, have not yet been eliminated, even in the most affluent societies. Basic needs concerning housing and food for all have not yet been secured, which makes health promotion meaningless for many people. We believe that a reduction in environmental exposures from pesticides, heavy metals, radon exposure, and other chemical and physical exposures are important, but at present these risk factors have a much smaller role than unhealthy lifestyle factors (Murray and Lopez 1999). As many of the non-communicable diseases develop over the entire lifespan prevention may be seen as a lifelong investment (Anonymous 1997). This enterprise should focus on a healthy diet (Glanz 1997), exercise (Gortmaker et al. 1997), and health promotion in general (Rijke 1997).
Changing lifestyle factors is more difficult and perhaps also more expensive than changing environmental factors. Not many people are willing to accept responsibility for their own health. Most would prefer to blame some external factor for their health problems—the environment in general, work conditions, or lack of social or personal support. Or they like to believe that medical treatment will solve all the problems that they may have in the future, perhaps not with the present technology but when they become patients in 10 to 20 years. The health-care industry has in many ways promised more than it could deliver. How much influence these promises have on health promotion and disease prevention is not known, but the detrimental effect could be substantial.
It is not impossible to change lifestyle, and many private enterprises achieve this through repeatedadvertising. Usually large budgets are required as it is often necessary to create a perceived need and then to maintainit. Some health-related changes definitively benefit from being linked to profit-making enterprises, for example improving physical fitness is supported by an industry selling sport clothes and sport equipment. Commercial interest in selling smoking cessation tools may also be much more efficient in reducing smoking habits than traditional anti-smoking campaigns. Therefore public effort should concentrate on preventing young people from starting to smoke as no companies have any commercial interest in this aspect. Although it is important to make the best use of private health-promoting activities it is difficult or even impossible to co-ordinate these private activities and to set priorities.
Setting up a system that provided safe drinking water needed financial resources and political leadership in Europe in the nineteenth century (Holland and Stewart 1998). The same is true for organizing a well-functioning health-care system with a strong emphasis on health promotion, but we do not really know what kind of political leadership is necessary to improve health behaviour. Nor do we know who has the necessary instruments to provide leadership of this scale. Health insurance companies, public health activities, private and public health care, pharmaceutical companies, strong medical professions, etc. do not have the same interests or goals, and they are powerful players in the health field and difficult to co-ordinate.
Prevention and care
Usually only a small fraction of the budget for health care is spent on prevention and most of the money is spent on screening and regular health examinations (Cohen and Henderson 1988), although most experts agree that only changes in the most important lifestyle factors will succeed in substantially improving global health indicators in developed countries, and in improving social conditions, family size, and safe water supply in developing countries. Despite this, most political attention, in both developed and developing countries, is focused on improving treatment. Furthermore, it is difficult to see how this could change. Treatment deals with named patients in need. Prevention is about anonymous individuals who are at present healthy and who, in general, will not know whethet or not they have benefited from the preventive action.
Convincing circumstantial evidence indicates that changing the sleeping position in early childhood from a prone to a supine position has saved thousands of children from sudden infant death syndrome (Taylor and Emery 1990; Taylor 1991; Fleming 1994), but we do not know who were saved and thus we have no grateful parents who donate money to research or tell their member of parliament that prevention is an important activity to support. Prevention may be better then cure. Still, prevention will never be able to compete with cure on resources for many reasons, some of which are reasonable. A utilitarian approach to health care would allocate more resources to disease prevention, but a strictly utilitarian approach is not acceptable for ethical reasons and will be in conflict with the aim of equity in treatment (Jensen and Mooney 1990). In like manner it will not be acceptable to shift resources from necessary and efficient treatments to prevention. On the other hand, many treatments have little or no scientific justification and in these situations money is much better spent on evidence-based prevention. However, there is little or no attempt to make this change in health policy. Only in secondary prevention in the form of screening is there political pressure for spending money on prevention that matches the pressure to support better treatment.
A health-care system needs to be organized towards well-defined health goals to make any difference in the priority setting. This message was probably the most important result of the WHO Health for All policy. Still, day-to-day problem solving, where the media’s treatment of case stories plays an important part, drives most health-care policy systems. In order to continue the discussion on setting priorities, the WHO (1999a) has decided to continue the Health for All strategy into the twenty-first century. A goal for chronic diseases has been set that aims at a 40 per cent reduction in mortality from cardiovascular diseases in people under 65, a 15 per cent reduction in cancer mortality under the age of 65, and a one-third reduction in diabetes-related amputations and pregnancy complications, and that 80 per cent of the children should be free of dental caries up to the age of 6. This goal is achievable even at our present level of knowledge.
While we do see political pressure to improve working conditions or the quality of drinking water, we cannot expect people to ask to be told what to do concerning their diet, smoking, and drinking habits. We may hope for support to make it easier to make healthy choices and for funding to provide information on the health consequences of the choices we make, but cannot demand much more than that.
Although prevention may be better than cure, few people act according to this precept. The health-care system is still organized to sweep up the water from an overflowing bathtub, and only a few people try to work out how to turn off the tap. It is unlikely that this will change substantially in the near future, which is why we should seize every opportunity to discuss and change unhealthy habits which do not provide personal satisfaction that justifies their negative health consequences. One such opportunity is illness. Stopping smoking after a myocardial infarction is late—too late in some cases, but not in all. Using secondary health-care institutions in prevention carries an important potential for lowering the disease burden. In most countries this potential resource has not been used to any large extent.
Reducing risk factors
Smoking is addictive. Addicts are strong advocates of their ‘drug’, whatever it may be, and smokers are no different from other addicts in this respect. Smokers may say that they smoke because they like smoking, but many like smoking because they dislike non-smoking. They have an urge for nicotine and fulfilling a need provides pleasure. Smokers often claim that they cannot think clearly without smoking; Sherlock Holmes described a difficult problem as a ‘three-pipe problem’. Smoking may of course stimulate brain activity, but the mechanism might as well be that low levels of nicotine slow down brain activity among the addicted. Smokers need not like smoking but they clearly dislike not smoking. Whatever the reasons for pleasure, all agree that smoking causes more serious health problems than any other single avoidable exposure (Doll and Peto 1981; Murray and Lopez 1999). It is likely that even passive smoking causes cancer and perhaps also coronary heart disease (Lam and He 1997). Therefore reducing smoking habits is part of most preventive programmes. The most effective smoking cessation methods are expensive and manpower consuming since they are based on individual counselling. Labelling tobacco products with health warnings, health campaigns, increased taxation on tobacco, and trying to make teachers and health-care workers understand that they act as role models are all important methods but are not sufficient in most countries. The WHO (1999b) estimates that the worldwide mortality from tobacco will rise from 4 million deaths per year in 1998 to 10 million in 2030. Half of these deaths will occur in people aged between 35 and 69 years. Smoking removes people not only from nursing homes but also from golf courses.
Use of nicotine substitutes has proved effective for highly motivated smokers if they are under close surveillance (Rose 1996). As a ‘stand-alone’ treatment without close surveillance the effect is less convincing. Only one study has tried to estimate the effect of buying nicotine patches over the counter without being offered additional help or surveillance, and it showed only a moderate effect in the subgroup that received the largest dose of nicotine (Sønderskov et al. 1997). Use of incentives in public campaigns has shown a very modest long-term effect. However, it is inexpensive and, even if only 1 per cent quits smoking after these campaigns (Bains et al. 1998), may be highly cost-effective compared with other methods.
The WHO (1999b) advocates four principles of tobacco control that have been successful in many countries: public health information, advertising bans, taxes, and building tobacco control coalitions. The public has a right to know about the health consequences of smoking, and labelling or counter-advertising may be used together with evidence-based information. Despite public campaigns, few people know all the health consequences of smoking and many believe that smoking only shortens life among the oldest. Higher taxes seem to reduce consumption, as so do smoke-free workplaces. Nicotine replacement therapy may be useful for many, especially if accompanied by other support systems. Tobacco control coalitions may counterbalance the tobacco industry lobbying and try to help people who depend on their income from tobacco sales to find other sources of income.
It is also well accepted that substantial modification of diet may have important health benefits. Doll and Peto (1981) estimated that 35 per cent of all cancers are related to diet, and an inappropriate diet is believed to cause obesity, diabetes, and cardiovascular diseases for some. Most believe that the amount of fat, especially saturated fat, is of importance, although trials have not shown convincing benefit of lowering fat in order to reduce the incidence of coronary heart diseases, perhaps because the reduction of fat intake has been too small. Results have been more promising in secondary prevention by reducing cardiovascular case fatality (Oliver 1996). It is also generally accepted that people eating a diet rich in fruits and vegetables have a low risk of cardiovascular diseases. Whether this is due to antioxidant vitamins is not known, but at present the advice is to eat fruit and vegetables rather than taking vitamin tablets or other antioxidants (Oliver 1996). A moderate intake of alcohol appears to lower the risk of coronary heart diseases (Rimm et al. 1999), but advocating a moderate alcohol intake may increase alcohol abuse.
A high-energy diet is probably suitable for humans performing heavy physical work. Most people in the developed countries no longer perform such work, and lack of exercise is a major health problem. Physical exercise appears to have a beneficial effect on our psychological well being by reducing depression, anxiety, and stress (Scully et al. 1998). Exercise may also lower blood pressure, perhaps by reducing weight (Arroll et al. 1994), cardiovascular diseases, osteoporosis, muscular disorders, and breast cancer (Deuster 1996; Pina and Fitzpatrick 1996; Batty 1997).
Social determinants of health
It is expected that disease and death would be closely correlated to poverty. It is disappointing that social inequalities in health are strong even in welfare states that have eradicated poverty in a materialistic sense for almost everyone in their societies and provided access to health care for all. It is perhaps difficult to avoid the fact that chronic disease can lead to unemployment and loss of income, but it should be within our reach to limit the inequalities in health that follow differences in social status. This has been one of the key targets in the WHO plan for Health for All by the Year 2000. However, in many countries the trend has been the opposite (Cavelaars et al. 1998). Reasons for social inequalities could be genetic, due to differences in access to health care, or due to differences in exposures to health hazards in the working environment or in personal life. Poor social conditions may also lead to greater exposure to stress and more changes in social conditions and other life events. It is unlikely that genetic factors alone or differential access to health care can explain more than a small fraction of the social differences in health we see in most countries. Social inequality exists even in countries with free and equal access to care, and genetic factors do not explain some of the rapidly changing social inequalities that we see in some countries. In some of the East European states there has been a substantial decline in life expectancy in males over the past 10 to 15 years (Leon et al. 1997). From 1970 to 1990 the social gradient in suicides has increased dramatically in some countries (Drever et al. 1996), and it is unlikely that genetic factors are the only reasons for this change. Better treatment of depression for some is probably a more likely explanation (Mortensen et al. 2000).
Changes in the social classification system may explain some of the changes in social inequalities over time. If the population is classified into, say, quintiles according to a given social indicator, selection bias hardly explains changes in social inequalities. However, if the social classification system is based upon educational levels, or some other classification system that changes over time, those who remain in the lower social groups need not be comparable with those from the same level in the past. This type of selection bias is expected to be present in many societies where social grouping is based on, for example, educational levels (Cavelaars et al. 1998).
Gender roles have changed over the years in many countries, mainly as a consequence of better education for females. How we best describe social conditions for the family in relation to health is not well studied, but results do indicate (Olsen and Frische 1993) that the social status of both partners should be taken into consideration, not only the one with the highest social level. It is reasonable to assume that the educational level plays a part for health behaviour regardless of the social status of the family. It is also of interest to follow social inequalities in health for men and women. Some results indicate that women do not always obtain the health benefits that their higher social positions indicate (DIKE 1997).
Social indicators are usually developed within social research. Although there are many possible ways that social factors may impair health (Marmot and Wilkinson 1999), we still need to know how best to classify social conditions in relation to health. Income, housing and working conditions, health behaviour, and access to health care may be related to health through very different mechanisms that may change over time and be different in different societies. How all these factors should be included in a social classification system that addresses health issues is not well studied and therefore is not known.
It is likely that mandatory public health programmes, or programmes offered to all, such as vaccination programmes (Holland 1995), free school meals, or control of work exposures help to reduce social inequalities in health. Health campaigns and voluntary screening programmes may, on the other hand, be better accepted by the best educated. Health-care workers have an important task in providing information to be used in primary or secondary prevention in a way that is understood by all. This potential for health improvements in patients with lifestyle-related diseases has not been widely used in most countries.
The mapping of the human genome provides new tools and new challenges in epidemiology (Schulte and Perera 1993). For most of the non-communicable diseases, the genetic risk factors will probably be complex and their individual contributions will be small. We hope that stratifying the participants on genetic factors will make it easier to identify environmental and preventable causes of diseases, as has been shown for the Leiden V mutation and oral contraception (Appleby and Olds 1997). However, other gene–environment interactions may be much more difficult to detect. Important genetic determinants of breast cancer, colorectal cancer, and cardiovascular diseases have been identified, but the importance of these findings in primary prevention is still not clear (Lander and Schork 1994; Ruiz et al. 1994; Olschwang et al. 1997; Duval et al. 1999).
Epidemiology has had a strong link to public health. New research methods incorporating molecular biology have the potential to enhance the ability of epidemiologists to study disease processes; however, it is important for those epidemiologists using these new molecular tools to use them to help elucidate public health issues (Olsen et al. 1999). The development of diseases over time cannot be understood outside a social context. Diseases have causes and many of these causes are man-made and, therefore, often avoidable. Although these causes interact with genetic factors to produce their effect, it seems more appropriate to target or prevent these causes rather than change susceptibility, which may have unknown side-effects.
Environmental risk factors
Since Ramazzini published his book De Morbis Artificium about occupational diseases in the eighteenth century, many diseases have been accepted as occupational or environmental. Percival Pott was the first to identify an occupational cancer in 1775 when he recognized soot from chimney sweeping as the cause of scrotal cancer. Much later the culprit was identified as one of the polycyclic hydrocarbons. Now we know that 4-aminophenyl, arsenic, asbestos, benzene, benzidine, chromium, polyclorinated biphenyls, vinyl chloride, and other compounds cause cancer. The fraction of cancers attributable to specific occupations is probably small in most developing countries. Most countries have been successful in finding substitutes for some of the carcinogens or in reducing exposure levels to very low levels, but heavy metals and pesticides with very long biological half-lives are of concern, especially if they accumulate in human food chains. Asbestos exposure and exposure to radon daughter elements are still widespread and constitute important public health problems (Last 1998).
Environmental diseases have to be identified and their determinants described. Some diseases are so closely related to their causes that the task may be easy, such as cancer of the nose in furniture workers or liver cell angiosarcoma in people exposed to vinyl chloride. Most other diseases have environmental as well as non-environmental causes and thus are more difficult to detect despite the fact that they are much more frequent. Weak associations between exposure to high levels of electromagnetic fields and childhood leukaemia (UK Childhood Cancer Study Investigators 1999) have been difficult to interpret, as weak associations indicate that other causes of the disease are not included in the statistical model used. We can only hope that these alternative causes are equally distributed among the exposed and non-exposed.
Once the exposures of importance are identified, they may be replaced or they may be isolated or not transmitted to people by means of efficient ventilation or use of personal protection.
In countries with very poor occupational standards education may be the most cost-effective way of reducing exposures. Reducing environmental exposures need not be very expensive if the exposure levels are high. Exposure to, say, organic solvents may often be greatly reduced by minor rearrangements at the work site. Knowing what to do is often the first step to getting it done, but many newly industrialized countries lack the knowledge or the infrastructure to implement the knowledge.
Unemployment, divorce, widowhood, and rapid welfare changes are all characterized by the disruption of social support systems and may be associated with poor health (Dean and Holstein 1991). Therefore disease prevention should also aim at making strong social networks when existing ones break down. Unions, churches, social clubs, schools, and sport clubs all have had this role. The health-care system also has a part to play for people who, for reasons of age, diseases, racial discrimination, etc., fall outside these systems. Providing social support to the elderly in Denmark showed remarkably reduced mortality and morbidity in one Danish study (Dean and Holstein 1991), although these results could only partly be replicated in another Danish study (Gunnar-Svensson et al. 1984). We do not know how much social support is needed to prevent diseases and we expect the support required to depend upon the amount and level of external stressors.
The philosophy behind tertiary prevention of chronic diseases is that it is often possible to live with and die with these diseases, rather than dying from them (Lorig et al. 1996). Research in general has shown that it is possible to prolong the time period of optimal, physical functioning, and social activity by providing social support and self-management programmes (Katz 1983; Breslow and Somers 1988).
A boring job with few social contacts may increase plasma adrenaline at rest and thus blood pressure. Other cardiovascular changes are seen for unemployment (Arnetz et al. 1991). Prevention at the work site is thus more than making sure that the workplace is safe; it also involves establishing a stimulating work environment and making sure that the workplace creates social contacts which do not interrupt the family networks. A safe job is a job that you like which also makes it possible to develop socially and intellectually.
A life-course approach to disease prevention
It is well known that, for example, neurotoxic exposures in early life may permanently impair brain functioning as is the case in fetal alcohol syndrome (Abel 1998). Infections may also cause damage that is not detected clinically within a short follow-up period, but could cause chronic disease in adulthood. Cardiovascular diseases, type II diabetes, and cancer have been seen as diseases that were the result of high-fat intake, smoking, and lack of physical exercise in adult life for genetically susceptible people. It is quite a recent development to consider this susceptibility to be not only a function of genetic factors, but also a result of exposures that took place early in life (Forsdahl 1977; Bakketeig et al. 1991; Barker 1994, 1995; Kuh and Ben-Shlomo 1997). It is known from animal experiments that the functioning of some organs may be permanently altered if the diet is poor in nutritional components such as proteins, and this alteration is called ‘programming’ (Eriksson and Swenne 1993; Lucas et al. 1999; Nieuwenhuizen et al. 1999). However, the idea that programming may also play a part for humans is still based on circumstantial evidence, although the evidence is strong. The first evidence came from Forsdahl’s (1977) studies in Norway showing that middle-aged males who were born in regions of high infant mortality at the time of their birth had a high mortality 50 years later, mainly due to cardiovascular diseases. Men born in regions with low infant mortality had a low risk of dying from cardiovascular diseases in middle-age. A similar ecological correlation was found to be present for serum cholesterol (Bakketeig et al. 1991). Barker (1995) followed up these ideas and established a link between birth weight and chronic diseases at the individual level in a series of important studies. In general, it was seen that low birth weight was associated with high blood pressure and high risk of other cardiovascular diseases in adult life, and the mechanism could operate via organ programming in fetal life (e.g. partly impaired liver functioning).
Low birth weight may be seen as a proxy measure of impaired fetal growth, although it is a very crude measure. Some neonates with low birth weight have grown to their full genetic potential, but others are growth retarded. If we are to look for fetal growth impairment at certain stages during pregnancy, it may be better to examine weight as a function of length (thinness), as it is done when estimating the Ponderal index (birth weight in kilograms divided by the cube of birth length in meters), which is, furthermore, a mathematical transformation that usually produces a symmetrical distribution of the index values. Birth weight and length are usually available in most birth records. On the other hand, in a search for growth retardation we should be looking for deviations from the birth weight that the fetus could have reached if it had access to use its full genetic growth potentials. Although this weight is unknown, we could calculate a predicted birth weight based on the gestational age and also on the parents’ heights or even their birth weights. If the child has siblings, their birth weights can be included in the model used for prediction. There are results that suggest that a deviation between the observed birth weight and the expected birth weight is a much better predictor of perinatal health problems than birth weight itself (Bakketeig and Hoffman 1983). Whether it is also a better correlate of organ programming, should it exist, remains to be seen.
In any case, there is now overwhelming evidence that birth weight correlates with several adult diseases and functions. Organ programming may indeed explain some of these associations, but confounding by genetic or social factors cannot be excluded. Birth weight could be an epiphenomenon of something causing low birth weight and the disease in question, and the two outcomes need not be causally related. Most likely programming will be the explanation for some of these associations, but hardly for all. The next critical step will be to see what initiates the lack of growth of certain cells in organ development. Many candidates are available; not only nutritional components but also smoking, work-related exposures, drugs, stress, alcohol, etc. may impair fetal growth. Diseases during pregnancy may in like manner impair fetal growth directly or via interference with fetal nutrition such as we expect to see in hyperemesis or conditions such as Crohn’s disease or ulcerative colitis (Fonager et al. 1998). Some of these growth-related factors may be preventable; others may not be. The major challenge in public health epidemiology is to identify those that are preventable and to use the information in the antenatal care system. Antenatal care will probably be an even more important part of disease prevention in the future than it is at present.
Although at present we do not know how important organ programming is for the most prevalent chronic diseases, most believe that it is of some importance. Unfortunately we do not know what activates the programming, but lack of nutrition has been part of most suggestions. In animal studies, specific nutrition components have been tested in most studies, especially a diet low in protein. In humans, lack of specific dietary components is also a possibility, although only components, such as n-3 fatty acids could explain the association seen for low birth weight and preterm delivery (S.F. Olsen 1993).
Diseases that could interfere with fetal growth or maternal nutrition, such as infections, hyperemesis, and placenta dysfunctions, could also play a part as well as external factors interfering with fetal growth, such as tobacco, alcohol, environmental exposures, or polluted drinking water or foods. We know that high levels of alcohol may permanently impair brain functioning, but probably not via programming (Vershuren 1993). We expect alcohol to be neurotoxic in high doses, and mercury may have a similar effect (Grandjean et al. 1999).
At present it is also believed that some cancers have a fetal aetiology, especially childhood cancers and cancer of the testis, but also breast cancers, ovarian cancers, and perhaps cancer of the prostate (Skakkebæk et al. 1987; Trichopoulos 1990; Ekbom et al. 1992, 1995; Adami et al. 1998; Ekbom 1998; Sabroe and Olsen 1998). The intrauterine hormonal level may be of importance for these cancers, especially oestrogen or the balance between oestrogen and progesterone, but again the evidence is circumstantial.
The number of Sertoli cells at birth determines sperm production, and male fecundity may similarly be affected by external or internal disrupters of the hormonal balance acting during the time period of organogenesis (Sharpe and Skakkebæk 1993). More than half of all pregnant women use medication during pregnancy. It is not known whether some types of medication have a programming effect or not.
However, whether or not these hypotheses are trueis not crucial to the life-course approach in preventing chronic diseases. We know that arteriosclerosis is a process that starts very early in life. We know that our dietary habits and our tendency to abuse alcohol, tobacco, or drugs depend upon social and psychological factors in our upbringing. We also know that obesity in childhood is strongly associated with obesity in adult life. Given these conditions, health promotion as well as disease management must focus on the longitudinal track of a given disease process on the health status of individuals as well as populations (Glassock 1997). Antenatal care is only the first process in a lifelong health promotion and disease prevention programme, and this programme needs to take into consideration not only diseases that surface to clinical detection shortly after the onset of the programme but also health in the long run. A health-care system strongly specialized within certain time periods of the lifespan or certain organ systems will not be well suited to meet the challenges raised by life-course research.
Non-communicable diseases in developing countries
Many developing countries are in a state of demographic balance with high fertility and a short life expectancy. The majority of the population are children or young adults. Improvements in social conditions usually decrease mortality and may increase fertility in the early stages (Taylor 1993), which leads to population growth before a new steady state is reached. This process is called the demographic transition, and it is usually followed by the epidemiological transition where the disease pattern changes from communicable to non-communicable diseases, although this pattern is not always seen. A deterioration of social conditions in industrialized countries in Eastern Europe has been followed not only by increased incidence of infectious diseases but also by increased mortality due to violent death and non-communicable diseases, especially cardiovascular diseases. This has been followed by a substantial reduction in life expectancy, especially for males (Leon et al. 1997).
The Inuit in Greenland have undergone a substantial improvement in social conditions that has not been followed by the expected improvement in health, despite a comprehensive health-care system that focuses on treatment rather than prevention. Social stress as a function of cultural changes has been followed by an increased consumption of unhealthy foods and frequent use of alcohol and tobacco. As in other rapid developing countries, blood pressure begins to rise with age and hypertension becomes a problem, which was not the case in the past. This change has been related to obesity, lack of exercise, and an increased salt intake. More psychosocial stress related to the cultural changes is probably also important (Bjerregaard and Young 1998).
Type II diabetes is a rapidly growing problem in many developing countries, especially in Pacific populations and among native people in Australia and the United States. The ‘thrifty genotype hypothesis’ (Neel 1999) states that the genetic susceptibility to type II diabetes improves survival during time periods of famine. Hypersecretion of insulin, which facilitates transfer of glucose into the cell, is an advantage when food is short in supply, but when food is too plentiful hypersecretion of insulin leads to pancreatic b-cell failure and diabetes.
Shortage of nutrition to the fetus may also lead to b-cell failure, which may lead to diabetes in adult life in case of obesity. Fetal programming may be an alternative explanation of the epidemic of diabetes seen in many populations throughout the world. In both situations, control over obesity by reducing the fat content in the diet and increasing exercise will reduce the incidence of diabetes. If fetal programming plays a part, additional preventive measures during pregnancy may present new preventive options once we know what activates the programming.
The epidemic of diabetes does have external and avoidable causes acting on biological susceptibility that may be due to genes or programming or both.
The epidemiological transition also changes the cancer pattern. Cancers which have an infectious aetiology, such as liver cancer, oral cancer, nasopharyngeal cancer, cervix cancer, and gastric cancer, are caused mainly by food components that carry the agent. These cancers tend to occur less frequently when social conditions improve. The cancers that take over are usually lung cancer if the tobacco companies are free to move into the market, breast cancer, and colorectal cancers. It is expected that changes in fertility partly explain the changes in breast cancer risk, although this has not been well documented. Changes in dietary patterns may explain the rise in the incidence of colon cancer.
In developing countries a health plan with emphasis on treatment and screening is bound to fail. Restriction of tobacco smoking, limiting of a high-fat Western diet, and promotion of physical activities in the young will be the only way to reduce premature death due to smoking and overeating. Unfortunately, many developing countries do not only take over ‘Western risk factors’ but also the industrialized society’s beliefs that better treatment in itself will solve all problems. However, the fight against AIDS has shown that prevention is possible, even in societies with a low level of background education. In many poor countries prevention is the only option available in fighting the disease burden of most importance.
Changes during the course of life
Critical changes occur in many countries which may permanently affect the health of a population. Individuals in all societies pass through critical stages during their life courses that may permanently determine future health. Social deprivation during childhood and lack of stimulation may not only lead to a low educational level and poverty, but may also have a direct biological affect on body development and disease susceptibility (Marmot and Wilkinson 1999).
Downward social mobility during a time period where the opposite is expected has been shown to be a stronger predictor for reproductive failures than low social status in itself (Basso et al. 1999). Social changes, like changes of schools, entry to the labour market, occupational changes, and retirement, may have positive or negative health effects depending on how well the individual copes with the psychological stress related to these events.
Unemployment is a risk factor for health in many situations, although research results have been ambiguous. Most studies have not been able to distinguish between health conditions leading to unemployment and health consequences of unemployment. Whether unemployment is a risk factor for health or not will depend on the social and psychological support provided to the unemployed as well as the working conditions present during employment. In some situations, unemployment will remove workers from high-risk exposures. We also expect the psychological stress to be less if many lose their job at the same time or if unemployment is common. Unemployment is probably worse for men if their roles in society are to be ‘bread winners’ in the family and if their social status depends upon their job. In any case, a policy that aims at reducing unemployment should also be supported for health reasons. Most research results indicate that unemployment may result in a deterioration in mental health, probably because a loss of job may be followed by a loss of a time structure to the day, a reduction in self-esteem, and lower social status.
Burden of chronic diseases
As indicated by the term ‘chronic’, these diseases do not show up in mortality statistics shortly after the onset and they need not lead to death. As only a few countries have reliable morbidity statistics, we do not know the global burden of chronic diseases in general, regardless of definition.
In order to get more information on chronic diseases that are not life threatening, several survey instruments have been developed. Most of these cover self-perceived physical and mental health problems leading to activity limitations (Hennessy et al. 1994; Lorig et al. 1996). In 1994 14 per cent of Danes reported health-related activity limitations. During a time period of 4 weeks 16 per cent reported reduced mental well being and 23 per cent said they had psychological problems leading to activity limitations. Poor health was reported more frequently in females than in males and poor health increased with age (Kjøller et al. 1999).
Murray and Lopez (1997a,b) have estimated the 30 leading causes of death worldwide and find that ischaemic heart diseases top the list, with cerebrovascular diseases in second place. Chronic obstructive pulmonary disease is sixth on the list, cirrhosis of the liver is thirteenth, and diabetes mellitus is sixteenth. In 1990 there were already 50 per cent more cancer deaths in less developed countries than in developed countries. Some diseases, such as diabetes, increase the risk of other diseases; compared with the estimated 580 000 deaths from diabetes in 1990, 2.8 million deaths were estimated to be attributable to diabetes. It is of interest that the probability of dying from a non-communicable disease was found to be higher in low-income regions such as Sub-Saharan Africa than in high-income regions (Murray and Lopez 1997b). If perinatal disease programming is a determinant of individual susceptibility, one can foresee a rapid growth in cardiovascular diseases in low-income regions following a higher life expectancy and better social conditions leading to sedentary work and a high-fat diet. This increase will exceed what is due to an increasing life expectancy and the age-adjusted predictions based upon current age- and sex-specific rates will underestimate the incidence.
While all accept that quality of life is at least as important as life expectancy in itself, it has been more difficult to agree upon an adjustment of the lost years of life for the quality aspect. Disability-adjusted life-years (DALYs) are now frequently used as a supplement to life expectancy measures (Murray and Lopez 1997c,d, 1999). DALYs are the sum of lost years of life due to death or disability adjusted for severity and as such are based upon value-loaded decisions. Murray and Lopez (1997b) estimated that in developed countries smoking is the leading cause of DALYs followed by alcohol, hypertension, and occupational risk factors. In developing countries malnutrition, poor water quality, unsafe sex, and alcohol were estimated to be the most important risk factors among the following: malnutrition, drinking water quality, unsafe sex, tobacco, alcohol, occupation, hypertension, physical inactivity, illicit drugs, and air pollution. However, they did not include overcrowding, unemployment, or social stress in their analyses. They did note that neuropsychiatric disorders were responsible for 10.5 per cent of the burden of diseases and injuries worldwide.
‘It is difficult to make predictions, especially for the future’ was a statement made by the Danish writer Storm P with which most people will agree. On the other hand, all preventive activities try to change undesirable expected future events and therefore are, or should be, based on predictions for the future. Most preventive activities address elimination or reduction of known risk factors, but often without giving much consideration to other consequences of behavioural or environmental changes.
Most of the present predictions are based on what we know about the long-term consequences of present risk factors or long-term trends in, for example, life expectancy. Based on information of this type we expect most populations to age in developed as well as in many developing countries and we expect non-communicable disease morbidity to increase from 28 million deaths worldwide in 1990 to 50 million in 2020. During the same time period we expect tobacco-related mortality to increase from 3 million to more than 8 million deaths. In 2020 the five leading causes of DALYs worldwide are expected to be ischaemic heart diseases, unipolar major depression, road traffic accidents, cerebrovascular diseases, and chronic obstructive pulmonary diseases (Murray and Lopez 1997b). Although we know how some of these diseases may be prevented, we do not in general know how to implement this knowledge, nor do we know the related consequences if we manage to eliminate or reduce these risk factors. Predicting future trends rests upon many assumptions, guesses, and intuition that call for the use of many different sources of information coming from experts as well as non-experts. These methods include Delphi techniques, brainstorming, and simulation games, as well as quantitative approaches in time-series analyses, use of Markov chains, and many other related techniques (Garrett 1999; Kirchhoff et al. 1999).
Although futures research is a risky matter, as time will show if we are right or wrong, it is necessary because it emphasizes the need to imagine what will happen with or without the preventive activity we want to implement. We have to try to capture all consequences of our preventive activities by focusing upon the larger picture rather than just single outcomes. For example, it is a noble aim to reduce childhood accidents in day-care centres, but as accidents are an unavoidable part of the way in which children explore and learn about the environment, they should not be reduced at all costs. Eradication of disease is extremely important for diseases such as smallpox, even though this makes the smallpox virus a frightening biological weapon. Elimination of measles will also be valuable if the virus is not replaced with other more harmful infectious agents, or if the disease or vaccination against it have no beneficial effect on the immune system in general. If the disease or the vaccination have a non-specific beneficial effect on the immune system, the net outcome may be different or perhaps even negative.
As diseases have causes, and often several causes, and as some of these causes are man-made and therefore subject to manipulation, studies of health futures need to specify the conditions for a prediction in order for it to be testable. By specifying these conditions, priority setting in public health work provides an important input to setting up work priorities. When these conditions are established, it becomes easier to predict the most likely future together with other possible health scenarios, including more unlikely but still possible scenarios. This is done to set the boundaries between possible and impossible outcomes.
In fact, ‘wild ideas’ should be encouraged in order to cover as many scenarios as possible. The next step could be to identify the most likely scenarioss. Methods such as the Delphi method or other expert-based surveys may be used to predict the most likely future based on the insight and information these experts possess.
In most research, consensus is reached by a critical scrutiny of the evidence. When all other possible explanations of an association have been disregarded, what remains could be the truth, as stated by Sherlock Holmes (Olsen et al. 1991). In futures research criticism need not be part of the process in setting up likely or possible scenarios.
Use of futures health studies on prevention is especially necessary when preventive activities interfere with other risk factors, people’s personal choice, or important social conditions. Use of hormonal replacement therapy not only prevents and causes a number of diseases, but also has an impact on sexual life and self-esteem. Taking such medicine regularly makes a healthy person a patient and has financial implications for women and perhaps also for society (Belchetz 1994).
Implementing a screening programme may bring more patients to treatment before their diseases reach a non-curable state, but may also delay treatment for those who are given a false-negative diagnosis. In adition, screening may induce anxiety among those who receive a false-positive result.
Screening programmes may replace primary prevention if it is believed that the health problem can be solved by early treatment. However, if a screening programme is followed by increased risk behaviour at work or in personal life, the net benefit may be less than expected and could even be negative.
Setting higher standards for automobile safety equipment leads to more expensive cars and may make it impossible for poor people to own a car, which could decrease job mobility and increase the risk of unemployment. Futures studies aim at bringing these problems into the open in order to minimize side-effects as much as possible by taking the appropriate actions in time.
The economy of prevention
Many countries declare that they want to give higher priorities to prevention in the future, but in general only little has been done in most countries. If priorities were set entirely upon how much health a unit of cost would produce, the health-care system would look very different from what we have at present. Few people would welcome such a radical approach as many other aspects have to be taken into consideration.
It is difficult to estimate how much money is actually spent on prevention as it will depend totally upon the definition of prevention. In the United Kingdom, attempts have been made to estimate the proportional distribution of the estimated 6 per cent spent on prevention in the National Health Service. Within the budget for prevention, it is found that 32.0 per cent is spent on the environment, 22.7 per cent on screening, 15.2 per cent on dental service, 7.7 per cent on maternal care, 7.6 per cent on occupational exposures, 5 per cent on family planning, 3.2 per cent on immunizations and vaccinations, and the remainder in other areas (Cohen and Henderson 1988).
Although it is likely that lasting improvements in health can secure considerable social and economic gains (WHO 1999b), priority setting should not be done on economic grounds only. Nor should prevention be implemented only if it saves resources elsewhere. We should accept that health in itself is a fair objective worth spending money on, and the decision process has to include intangibles and aim at equity.
Chronic diseases will increase in numbers when life expectancy is prolonged due to better social conditions, less crowding (Aaby and Samb 1994), better treatment of acute infections, and active immunization. However, chronic diseases may be delayed and the time period of active life prolonged. The length of good-quality life is what should concern us, not whether or not we contract the disease. This idea is accepted by most when we talk about death.
As prevention aims at improving true quality of life—to add life to years, rather than adding years to life—we should respect people’s choices when selecting personal habits. We may be firmer when hazards that give no pleasure are to be removed from the environment, or when we try to make sure that everyone is able to live a socially acceptable life, to have a job, and to have access to education that is appropriate for his or her personal capacity and aspirations. Chronic disease must be prevented both within and outside the health-care sector. Most countries lack a high-ranking advocate for this activity—perhaps the time has come to have ministers for health promotion.
Aaby, P. and Samb, B. (1994). In Health and diseases in developing countries (ed. K.S. Lankinen, S. Bergström, P.H. Mäkelä, and M. Peltoman), pp. 166–76. Macmillan, London.
Abel, E.L. (1998). Prevention of alcohol abuse-related birth effects. II: Targeting and pricing. Alcohol and Alcoholism, 33, 417–20.
Adami, H.O., Signorello, L.B., and Trichopoulos, D. (1998). Towards an understanding of breast cancer etiology. Seminars in Cancer Biology, 8, 255–62.
Anonymous (1997). Guidelines for school health programs to promote lifelong healthy eating. Journal of School Health, 67, 9–26.
Appleby, R.D. and Olds, R.J. (1997). The inherited basis of venous thrombosis. Pathology, 29, 341–7.
Arnetz, B.B., Brenner, S.O., Levi, L., et al. (1991). Neuroendocrine and immunologic effects of unemployment and job insecurity. Psychotherapy and Psychosomatics, 55, 76–80.
Arroll, B., Hill, D., White, G., Sharpe, N., and Beaglehole, R. (1994). The effect of exercise episode duration on blood pressure. Journal of Hypertension, 12, 1413–15.
Bains, N., Pickett, W., and Hoey, J. (1998). The use and impact of incentives in population-based smoking cessation programs: a review. American Journal of Health Promotion, 12, 307–20.
Bakketeig, L.S. and Hoffman, H.J. (1983). The tendency to repeat gestational age and birth weight in successive births, related to perinatal survival. Acta Obstetricia et Gynecologica Scandinavica, 62, 385–92.
Bakketeig, L.S., Magnus, P., and Sundet, J.M. (1991). In Problems and methods in longitudinal research, (ed. D. Magnusson, L.R. Bergman, G. Rudinger, and B. Törestad). Cambridge University Press.
Barker, D.J.P. (1994). Mothers, babies and disease in later life. BMJ Publishing, London.
Barker, D.J. (1995). Fetal origins of coronary heart disease. British Medical Journal, 311, 171–4.
Basso, O., Olsen, J., and Christensen, R. (1999). Study of environmental, social, and paternal factors in preterm delivery using sibs and half sibs. Journal of Epidemiology and Community Health, 53, 20–3.
Batty, D. (1997). Review of interventions to prevent heart disease. British Medical Journal, 315, 1468.
Belchetz, P.E. (1994). Hormonal replacement in postmenopausal women. New England Journal of Medicine, 330, 1062–71.
Berry, R.J., Li, Z., Erickson, J.D., et al. (1999). Prevention of neural-tube defects with folic acid in China. New England Journal of Medicine, 341, 1485–90.
Bjerregaard, P. and Young, T.K. (1998). The circumpolar Inuit: health of a population in transition. Munksgaard, Copenhagen.
Breslow, L. and Somers, A. (1988). The periodic health examination, and updates. Canadian Medical Association Journal, 4, 617–26.
Brønnum-Hansen, H. (1999). How good is the Prevent model for estimating the health benefits of prevention? Journal of Epidemiology and Community Health, 53, 300–5.
Cavelaars, A.E.J.M., Kunst, A.E., Geurts, J.J., et al. (1998). Differences in self reported morbidity by educational level: a comparison of 11 Western European countries. Journal of Epidemiology and Community Health, 52, 219–27.
Cohen, D.R. and Henderson, J.B. (1988). Health, prevention and economics. Oxford University Press.
van Dam, J., Bond, J.H., and Sivak, M.V. Jr. (1995). Fecal occult blood screening for colorectal cancer. Archives of Internal Medicine, 155, 2389–402.
Dean, K. and Holstein, B.E. (1991). In Health promotion research (ed. B. Bandura and I. Kickbusch). WHO European Series 37. WHO, Copenhagen.
De Koning, H.J. (2000). Breast cancer screening: cost-effective in practice? European Journal of Radiology, 33, 32–7.
Deuster, P.A. (1996). Exercise in the prevention and treatment of chronic disorders. Women’s Health Issues, 6, 320–31.
DIKE (1997). Danskernes sundhed mod år 2000. Sundhedsadfærd, sundhedstilstand, sygelighed, dødelighed, levekår. DIKE, Copenhagen.
Doll, R. and Peto, R. (1981). The causes of cancer. Quantitative estimates of avoidable risks of cancer in the United States today. Oxford University Press.
Drever, F., Whitehead, M., and Roden, M. (1996). Current patterns and trends in male mortality by social class. Population Trends, 86, 15–20.
Duval, A., Gayet, J., Zhou, X.P., Iacopetta, B., Thomas, G., and Hamelin, R. (1999). Frequent frameshift mutations of the TCF-4 gene in colorectal cancers with microsatellite instability. Cancer Research, 59, 4213–15.
Ekbom, A. (1998). Growing evidence that several human cancers may originate in utero. Seminars in Cancer Biology, 8, 237–44.
Ekbom, A., Trichopoulos, D., Adami, H.O., Hsieh, C-C., and Lan, S-J. (1992). Evidence of prenatal influences on breast cancer risk. Lancet, 340, 1015–18.
Ekbom, A., Thurfjell, E., Hsieh, C.C., Trichopoulos, D., and Adami, H.O. (1995). Perinatal characteristics and adult mammographic patterns. International Journal of Cancer, 61, 177–80.
Eriksson, U.J. and Swenne, I. (1993). Diabetes in pregnancy: fetal macrosomia, hyperinsulinism, and islet hyperplasia in the offspring of rats subjected to temporary protein-energy malnutrition early in life. Pediatric Research, 34, 791–5.
Executive Board of the WHO (2000). Prevention and control of noncommunicable diseases. EB105.R12, 27 January 2000. WHO, Geneva.
Fleming, P.J. (1994). Understanding and preventing sudden infant death syndrome. Current Opinion in Pediatrics, 6, 158–62.
Fonager, K., Sørensen, H.T., Olsen, J., Dahlerup, J.F., and Rasmussen, S.N. (1998). Pregnancy outcome for women with Crohn’s disease: a follow-up study based on linkage between national registries. American Journal of Gastroenterology, 93, 2426–30.
Forsdahl, A. (1977). Are poor living conditions in childhood and adolescence an important risk factor for arteriosclerotic heart disease? British Journal of Preventive and Social Medicine, 31, 91–5.
Garrett, M.J. (1999). Health futures. WHO, Geneva.
Glanz, K. (1997). Dietary change. Cancer Causes and Control, 8, S13–16.
Glassock, R.J. (1997). Optimizing disease management in the next 25 years. Seminars in Nephrology, 17, 387–90.
Gortmaker, S.L., Mariani, A., Peterson, K., Cheung, L., and Wiecha, J. (1997). Exercise. Cancer Causes and Control, 8, S17–19.
Gotzsche, P.C. and Olsen, O. (2000). Is screening for breast cancer with mammography justifiable? Lancet, 355, 129–34.
Grandjean, P., Budtz-Jorgensen, E., White, R.F., et al. (1999). Methylmercury exposure biomarkers as indicators of neurotoxicity in children aged 7 years. American Journal of Epidemiology, 150, 301–5.
Gunnar-Svensson, F., Ipsen, J., Olsen, J., and Waldstrøm, B. (1984). Prevention of relocation of the aged in nursing homes. Scandinavian Journal of Primary Health Care, 2, 49–56.
Gyrd-Hansen, D. (1999). The relative economics of screening for colorectal cancer, breast cancer and cervical cancer. Critical Reviews in Oncology/Hematology, 32, 133–44.
Hedner, T. (1998). Treating hypertension—effect of treatment and cost-effectiveness in respect to later cardiovascular diseases. Scandinavian Cardiovascular Journal, Supplement, 47, 31–5.
Hennessy, C.H., Moriarty, D.G., Zack, M.M., Scherr, P.A., and Brackbill, R. (1994). Measuring health-related quality of life for public health surveillance. Public Health Reports, 109, 665–72.
Holland, W.W. (1995). Advances in epidemiology and disease prevention. Annals of the Academy of Medicine, Singapore, 24, 230–7.
Holland, W.W. and Stewart, S. (1998). Public health. The vision and the challenge. Nuffield Trust, London.
Holloway, R.G., Benesch, C.G., Rahilly, C.R., and Courtright, C.E. (1999). A systematic review of cost-effectiveness research of stroke evaluation and treatment. Stroke, 30, 1340–9.
Hughes, K. (1997). Screening for and treatment of hypercholesterolaemia: a review. Annals of the Academy of Medicine, Singapore, 26, 215–20.
Hugod, C. and Fog, J. (ed.) (1992). Screening. Why, when and how? National Board of Health, Denmark (English translation).
Jensen, U.J. and Mooney, G. (1990). Changing values in medical and health care decision making. Wiley, Chichester.
Jatoi, I. (1999). Breast cancer screening. American Journal of Surgery, 177, 518–24.
Katz, S. (1983). Active life expectancy. New England Journal of Medicine, 11, 1218–24.
Kirchhoff, M., Davidsen, M., Bronnun-Hansen, H., et al. (1999). Incidence of myocardial infarction in the Danish MONICA population 1982–1991. International Journal of Epidemiology, 28, 211–18.
Kjøller, M., Rasmussen, N.K., and Keiding, L.M. (1999). Selvrapporteret sundhed og sygelighed blandt voksne danskere 1987–1994. Ugeskrift for Laeger, 16, 2948–54.
Koplan, J.P. and Livengood, J.R. (1994). The influence of changing demographic patterns on our health promotion priorities. American Journal of Preventive Medicine, 10, 42–4.
Kronborg, O., Fenger, C., Olsen, J., Jorgensen, O.D., and Søndergaard, O. (1996). Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet, 348, 1467–71.
Kuh, D. and Ben-Shlomo, B. (ed.) (1997). A life course approach to chronic disease epidemiology. Tracing the origins of ill-health from early to adult life. Oxford University Press.
Lam, T.H. and He, Y. (1997). Passive smoking and coronary heart disease: a brief review. Clinical and Experimental Pharmacology and Physiology, 24, 993–6.
Lander, E.S. and Schork, N.J. (1994). Genetic dissection of complex traits. Science, 265, 2037–48.
Last, J.M. (1998). Public health and human ecology. Appleton and Lange, Stamford, CT.
Leon, D.A, Chenet, L., Shkolnikov, V.M., et al. (1997). Huge variation in Russian mortality rates 1984–94: artefact, alcohol, or what? Lancet, 350, 383–8.
Littenberg, B. (1995). A practice guideline revisited: screening for hypertension. Annals of Internal Medicine, 122, 937–9.
Lorig, K., Stewart, A., Ritter, P., González, V., Laurent, D., and Lynch, J. (1996). Outcome measures for health education and other health care interventions. Sage Publications, London.
Lucas, A., Fewtrell, M.S., and Cole, T.J. (1999). Fetal origins of adult disease—the hypothesis revisited. British Medical Journal, 319, 245–9.
Lyon, A. (1967). Causality. British Journal of the Philosophy of Science, 1–20.
McKeown, T. (1965). Medicine in modern society. Medical planning based on evaluation of medical achievement. George Allen & Unwin, London.
McKeown, T. and Lowe, C.R. (1974). An introduction to social medicine. Blackwell Scientific, Oxford.
McMichael, A.J. (1999). Prisoners of the proximate: loosening the constraints on epidemiology in an age of change. American Journal of Epidemiology, 149, 887–97.
Mackie, J.L. (1974). The cement of the universe. Oxford University Press.
Marmot, M. and Wilkinson, R.G. (1999). Social determinants of health. Oxford University Press.
Mortensen, P.B., Agerbo, E., Erikson, T., Qin, P., and Westergaard-Nielsen, N. (2000). Psychiatric illness and risk factors for suicide in Denmark. Lancet, 355, 9–12.
Murray, C.J.L. and Lopez, A.D. (1997a). Alternative projections of mortality and disability by cause 1990–2020: Global Burden of Disease Study. Lancet, 349, 1498–504.
Murray, C.J.L. and Lopez, A.D. (1997b). Global mortality, disability, and the contribution of risk factors: Global Burden of Disease Study. Lancet, 349, 1436–42.
Murray, C.J.L. and Lopez, A.D. (1997c). Mortality by cause for eight regions of the world: Global Burden of Disease Study. Lancet, 349, 1269–76.
Murray, C.J.L. and Lopez, A.D. (1997d). The utility of DALYs for public health policy and research: a reply. Bulletin of the World Health Organization, 75, 577–81.
Murray, C.J.L. and Lopez, A.D. (1999). On the comparable quantification of health risks: lessons from the Global Burden of Disease Study. Epidemiology, 10, 594–605.
Neel, J.V. (1999). Diabetes mellitus: a ‘thrifty’ genotype rendered detrimental by ‘progress’. Bulletin of the World Health Organization, 77, 694–703.
Nieuwenhuizen, A.G., Schuiling, G.A., Seijsener, A.F., Moes, H., and Koiter, T.R. (1999). Effects of food restriction on glucose tolerance, insulin secretion, and islet-cell proliferation in pregnant rats. Physiology and Behavior, 65, 671–7.
Oliver, M.F. (1996). Which changes in diet prevent coronary heart disease? A review of clinical trials of dietary fats and antioxidants. Acta Cardiologica, 6, 467–90.
Olsen, J. (1993). Some consequences of adopting a conditional deterministic causal model in epidemiology. European Journal of Public Health, 3, 204–9.
Olsen, J. and Frische, G. (1993). Social differences in reproductive health. A study on birth weight, still births and congenital malformations in Denmark. Scandinavian Journal of Social Medicine, 2, 90–7.
Olsen, J., Merletti, F., Snashall, D., and Vuylsteek, K. (1991). Searching for causes of work-related diseases. An introduction to epidemiology at the work site. Oxford University Press.
Olsen, J., Andersen, P.K., Sørensen, T.I.A., Melbye, M., and Aaby, P. (1999). The future of epidemiology. International Journal of Epidemiology, 28, S996.
Olsen, S.F. (1993). Marine n-3 fatty acids ingested in pregnancy as a possible determinant of birth weight: a review of the current epidemiologic evidence. Epidemiologic Reviews, 15, 399–413. (Erratum. American Journal of Epidemiology, 139, 856 (1994).)
Olschwang, S., Hamelin, R., Laurent-Puig, P., et al. (1997). Alternative genetic pathways in colorectal carcinogenesis. Proceedings of the National Academy of Sciences of the United States of America, 94, 12122–7.
Pina, I.L. and Fitzpatrick, J.T. (1996). Exercise and heart failure. Chest, 110, 1317–27.
Rijke, R.P. (1997). Health promotion from an ecological viewpoint. Oncologica, 14, 34–6 (in Dutch).
Rimm, E.B., Williams, P., Fosher, K., Criqui, M., and Stampfer, M.J. (1999). Moderate alcohol intake and lower risk of coronary heart disease: meta-analysis of effects on lipids and haemostatic factors. British Medical Journal, 319, 1523–8.
Roemer, M.I. (1984). The value of medical care for health promotion. American Journal of Public Health, 74, 243–8.
Rose, G. (1992). The strategy of preventive medicine. Oxford University Press.
Rose, J.E. (1996). Nicotine addiction and treatment. Annual Review of Medicine, 47, 493–507.
Rothman, K.J. (1976). Causes. American Journal of Epidemiology, 104, 587–92.
Ruiz, J., Blanche, H., Cohen, N., et al. (1994). Insertion/deletion polymorphism of the angiotensin-converting enzyme gene is strongly associated with coronary heart disease in non-insulin-dependent diabetes mellitus. Proceedings of the National Academy of Sciences of the United States of America, 91, 3662–5.
Sabroe, S. and Olsen, J. (1998). Perinatal correlates of specific histological types of testicular cancer in patients below 35 years of age: a case-cohort study based on midwives’ records in Denmark. International Journal of Cancer, 78, 140–3.
Schueler, K. (1994). Cost-effectiveness issues in hypertension control. Canadian Journal of Public Health, 85, S54–6.
Scully, D., Kremer, J., Meade, M.M., Graham, R., and Dudgeon, K. (1998). Physical exercise and psychological well being: a critical review. British Journal of Sports Medicine, 32, 111–20.
Schulte, P.A. and Perera, F.P. (1993). Molecular epidemiology: principles and practice. Academic Press, San Diego, CA.
Sharpe, R.M. and Skakkebæk, N.E. (1993). Are oestrogens involved in falling sperm counts and disorders of the male reproductive tract. Lancet, 341, 1392–5.
Skakkebæk, N.E., Berthelsen, J.G., Giwercman, A., and Müller, J. (1987). Carcinoma-in-situ of the testis: possible origin from gonocytes and precursor of all types of germ cell tumours except spermatocytoma. International Journal of Andrology, 10, 19–28.
Sønderskov, J., Olsen, J., Sabroe, S., Meillier, L., and Overvad, K. (1997). Nicotine patches in smoking cessation: a randomized trial among over-the-counter customers in Denmark. American Journal of Epidemiology, 145, 309–18.
Taylor, B.J. (1991). A review of epidemiological studies of sudden infant death syndrome in southern New Zealand. Journal of Paediatrics and Child Health, 27, 344–8.
Taylor, E.M. and Emery, J.L. (1990). Categories of preventable unexpected infant deaths. Archives of Disease in Childhood, 65, 535–9.
Taylor, R. (1993). Non-communicable diseases in the tropics. Medical Journal of Australia, 159, 266–70.
Trichopoulos, D. (1990). Hypothesis: does breast cancer originate in utero? Lancet, 335, 939–40.
UK Childhood Cancer Study Investigators (1999). Exposure to power-frequency magnetic fields and the risk of childhood cancer. Lancet, 354, 1925–31.
Vershuren, P.M. (ed.) (1993). Health issues related to alcohol consumption. ILSI Europe, Brussels.
WHO (1999a). Health 21. The health for all policy framework for the WHO European Region. European Health for All Series 6, WHO, Copenhagen.
WHO (1999b). The world health report 1999. Making a difference. WHO, Geneva.
Wynder, E.L. and Andres, R. (1994). Workshop A: diet and nutrition research as it relates to aging and chronic diseases. Preventive Medicine, 23, 549–51.