Socioeconomic inequalities in health in the Netherlands: impact of a five year research programmeBMJ 1994; 309 doi: https://doi.org/10.1136/bmj.309.6967.1487 (Published 03 December 1994) Cite this as: BMJ 1994;309:1487
- Johan P Mackenbach, professor of public healtha
- a Department of Public Health, Erasmus University, Rotterdam, PO Box 1738, 3000 DR Rotterdam, Netherlands
- Accepted 7 November 1994
The attention paid to the socioeconomic inequalities in health in the Netherlands has increased greatly in recent years. A national research programme was started in 1989, and among other things, this has increased the yearly number of publications on socioeconomic inequalities in health by about 25%. The programme has increased awareness of inequalities among researchers and policy makers as well as improved the information available on health inequalities and the reasons for them. Cross party agreement on the need to reduce these inequalities has led to a consensus based approach which contrasts with the heavily politicised debate in countries such as the United Kingdom.
The publication of the Black report in 19801 and the inclusion of a reduction of health inequalities among the World Health Organisation's Health for All policy targets in 19852 has increased interest in socioeconomic inequalities in health in many countries. In the Netherlands, socioeconomic inequalities in health were politically a non-issue until the second half of the 1980s. In 1980 the Dutch Society for Social Medicine celebrated its 50th anniversary with a conference on socioeconomic inequalities in health,3 and the results of a thorough study of inequalities between neighbourhoods in Amsterdam were published,4 but neither of these initiatives was given political follow up. The position changed radically, however, after the Dutch government adopted the Health For All policy targets, and in 1986 the ministry of welfare, public health, and cultural affairs included a paragraph on inequalities in health in an important policy document.5
After that, initiatives were taken from inside the ministry to put equity in health on the political agenda.6 In 1987 a conference was organised under the aegis of the prestigious Scientific Council for Government Policy, and widespread press coverage was arranged for the publication of its proceedings. The report contained a proposal for a national research programme which was launched in 1989. An independent committee was formed to develop and implement the research programme, and to report the results to the minister of welfare, public health, and cultural affairs after five years. The programme had an annual budget of 1 million fl (about £370000), which could be spent at the discretion of the committee. The committee consisted of an independent chairman, established researchers, a representative for the ministry, and several people familiar with (but not representing) other policy areas. The main objectives of the programme were to generate more knowledge about the size and nature of socioeconomic inequalities in health and the reasons for them. The results have been published recently.7 8
Effects on the scientific community
The programme was designed so that its effect on the scientific community would last longer than five years. As many research groups as possible were therefore involved in the programme. Forty studies were commissioned, most of which were small scale secondary analyses of data collected in epidemiological studies. This was done to increase awareness of socioeconomic inqualities in health among researchers working in other areas.
At the same time, investments were made to improve conditions for future research. For example, a standardised procedure for measuring socioeconomic status on the basis of education, occupational class, or income was developed to improve the comparability of research findings.9 A computerised method for eliciting this information from patients admitted to a hospital was also developed, and tests showed that this method will allow data on socioeconomic status to be incorporated into all kinds of routine data collections.10
Although it is too early to know whether the research programme has had a lasting effect on the scientific community, it has had a measurable effect in the short term. A documentation centre set up to monitor the scientific literature on socioeconomic inequalities in health from the Netherlands and abroad showed that the number of Dutch publications increased by about 25% during the programme (table I). The centre has a virtually complete coverage of articles published in Dutch and English language peer reviewed journals since 1985. In the United Kingdom the yearly number of publications fell slightly during the same period, suggesting that the increase observed in the Netherlands is not simply the effect of an international trend. Some of the larger independent research funds in the Netherlands, such as the Sick Fund Council (Ziekenfondsraad)11 and the Netherlands Heart Foundation (Nederlandse Hartstichtung),12 have sponsored studies on health inequalities, and the Prevention Fund (Praeventiefonds) has selected socioeconomic inqualities in health as one of its top research priorities for 1994-7.13
Increased knowledge of inequalities
When the research programme started the Netherlands had few data describing the size and nature of socioeconomic inequalities in health. A review for the conference of the Scientific Council for Government Policy in 1987 had to fill in many gaps in knowledge by referring to studies from neighbouring countries.14 Reviews in 1992 and this year were able to offer a much more complete picture of the Dutch situation.8 15 This is because of more documenting of socioeconomic inequalities in health by the Netherlands Central Bureau of Statistics and the large number of secondary analyses of epidemiological data.
In contrast with many other European countries the Netherlands does not have mortality statistics by socioeconomic status, but studies relating mortality of city neighbourhoods or regions to socioeconomic characteristics have shown that mortality is higher in lower socioeconomic groups.4 16 These findings are confirmed by the results of several epidemiological studies of specific cohorts whose mortality could be analysed according to socioeconomic indicators at the start of follow up.17 18 19 20
The most important source of continuous data on socioeconomic inqualities in health is the Netherlands health interview survey. This survey is conducted by the Central Bureau of Statistics and has been used extensively during the research programme.21 22 In almost all its publications the bureau now presents breakdowns by level of education and income, which has greatly increased the public visibility of socioeconomic inequalities in health. All health indicators measured in this survey (perceived general health; subjective health complaints; reported chronic conditions; reported physical disabilities, etc) show large variations with level of education and income.21 22 Of course, differences in reporting by socioeconomic status could bias these results, but the available evidence on the accuracy of reporting of chronic conditions in Dutch health surveys suggests that if such bias exists it will lead to underestimation of socioeconomic inequalities.23 24
The box summarises the descriptive evidence on socioeconomic inequalities in health in the Netherlands. Data were collected from the health interview survey, cancer registries (incidence25 and survival26 of cancer); epidemiological studies of cardiovascular disease,27 chronic obstructive lung disease,24 tooth decay,28 and psychiatric disorders29; and registries of sickness absence and long term work disability.30 Boshuizen et al combined data on socioeconomic variations in mortality and self reported morbidity to estimate variations in healthy life expectancy and found that Dutch men with higher education have 12 years longer healthy life expectancy than men with lower education.31 The substantial socioeconomic inequalities in health in the Netherlands can also be shown by population attributable risks (table II), which show that if people with lower education had the morbidity and mortality of those with university education the average morbidity and mortality in the Dutch population would be reduced by 25-50%. Nevertheless inequalities in health are still smaller in the Netherlands than in most other industrialised countries.*RF 32-34* For example, among young adult men the difference in mortality between those with the lowest and those with the highest educational level is about 70% in the Netherlands, 100% in England and Wales, 200% in France, and 260% in the United States.34
Health indicators that are more common in lower socioeconomic groups8
Self reported health problems:
Less than “good” general health
Subjective health complaints
Chronic obstructive lung disese
Low back pain
Arthritis and arthrosis
Psychological and psychosocial problems
Having no teeth
Health problems identified through registries or medical examinations:
Chronic obstructive lung disease
Dental caries and periodontal disorders
Short term work disability
Long term work disability
Short survival from cancer
Increased knowledge of causes of inequality
Although the causes of socioeconomic inequalities in health are partly known in other countries, this information is not necessarily applicable to the Dutch situation because, for example, the socioeconomic distribution of risk factors for disease might be different in the Netherlands from that in other countries. A substantial part of the budget of the research programme in health was therefore devoted to explaining socioeconomic inequalities in the Netherlands. Table III summarises the available evidence. Considerable progress has been made, especially with regard to the effect of life style factors, working and housing conditions, and psychosocial factors. Many of the well known determinants of disease occur more often in the lower socioeconomic groups (figure).
Use of preventive and curative health services is not consistently associated with socioeconomic status in the Netherlands. After variations in health are controlled for some services are used more often by people with higher educational levels (for example, influenza vaccinations,35 outpatient specialist and physiotherapy services, and ambulatory mental health services22) but others are used more often by people with lower educational levels (for example, general practitioner services22) or show no differences at all.
The main task, however, is not to document inequalities in determinants of health problems but to estimate the contribution of each of these determinants to the health inequalities. Multivariate analyses of the simultaneous relation between socioeconomic status, specific health determinants, and frequency of health problems, have until now been limited to a few examples*RF 36-39* and suffer from various limitations. The results are summarised in the second column of table III. One of the most interesting studies attempted to explain socioeconomic variation in the prevalence of heart disease from variation in the prevalence of classic risk factors such as smoking, hypertension, and high serum cholesterol concentration. About one third of the differences in prevalence of heart disease could be explained by these risk factors.36 This finding is similar to the main results of the British Whitehall study, which suggested that other factors also need to be considered.40
More work needs to be done before we have a full understanding of the causes of socioeconomic inequalities in health. For this reason a large scale longitudinal study has been started, sponsored by the research programme.41
Effects on health policy
Although it is too early to have a good picture of the changes in health policy induced by the results of the research programme, there are some encouraging signs that socioeconomic inequalities in health are being included in health policy. The Scientific Council for Government Policy held a second conference in 1991 entitled “Socioeconomic health inequalities and policy.” Representatives from most political parties, from relevant ministries such as social affairs and housing, from health care organisations, and the medical professions discussed potential interventions to reduce socioeconomic inequalities in health. There was a broad consensus that these inequalities are unfair, and that all those involved should try to contribute to redressing the inequalities.42
Since this conference, several initiatives have been taken, at national, regional, and local levels. At the national level, an intersectoral working group was formed to stimulate cooperation between the various ministries. At the regional and local levels, many public health departments have intensified their efforts at improving health related living circumstances in deprived areas. The international Healthy Cities movement and a government policy aiming at social renewal have both played a part, while the good data on socioeconomic inequalities in health have helped to give these initiatives a clear focus. Many towns are experimenting with intervention programmes in deprived areas, encompassing such diverse elements as increased safety from violence, urban renewal, health education campaigns, and help with finding jobs.
A new five year research programme is starting this year to develop and evaluate community interventions to reduce health problems in lower socioeconomic groups. This programme will be based on the recommendations issued by the committee overseeing the 1989-93 research programme. In its final report to the deputy minister of welfare, public health, and cultural affairs the committee gave several recommendations on the basis of an explanation of socioeconomic inequalities in health in terms of causation (through the differential distribution across socioeconomic groups of specific health determinants such as smoking, working conditions, and psychosocial stress) and in terms of selection (through an effect of health on social mobility).7 Almost no data on selection are available for the Netherlands, and international data suggest that it is less important than causation,43 but it was thought important to identify all possible routes of intervention. Four types of intervention were identified (see below), and these will form the basis for the research and development efforts of the 1994-8 programme.
Improving the educational, occupational, or income level of those at the bottom of the social hierarchy. This is the most fundamental approach to diminishing the excess morbidity and mortality among those with lower socioeconomic status and perhaps, therefore, the potentially most effective approach. But it can be implemented only in the longer term, and the current political climate presents serious barriers especially for redressing income inequalities. The recommendations therefore were limited to raising the awareness among policy makers of the implications that their decisions could have for the health of the population. One specific recommendation was to assess the possible health effects, especially among the disadvantaged, of all important policy changes in education, employment, and income (including social security).
Minimising the effects of ill health on social mobility. Although this may be a less important mechanism for the overall explanation of socioeconomic inequalities in health, it offers opportunities for intervention that can be built into existing services and policies. Sick children who perform less well in school and their teachers could receive extra guidance and help from school health services. Chronically ill adults who have difficulty in finding or keeping a job could receive extra help from employers and receive decent income supplements when they are unemployed.
Reducing exposure to determinants of health problems in the lower socioeconomic groups. This is a feasible and potentially powerful way of redressing socioeconomic inequalities in health. Unhealthy lifestyles and unhealthy working conditions are the top priorities for action. Unhealthy lifestyles cannot be changed simply by giving more conventional health education. Methods of health education need to be developed that take into account the specific characteristics of the target groups and should be combined with efforts to change the incentive structure for these behaviours among disadvantaged groups of the population. In campaigns aimed at improving working conditions extra attention could be paid to the less well paid jobs.
Offering extra health care to lower socioeconomic groups. This is the least fundamental approach to reducing socioeconomic inequalities in health but it may be effective, especially if the other approaches have not been completely successful (as is likely to be the case) and if the health services offered have the potential of really alleviating the health problems. One example may be general practitioner care in deprived areas. Efforts at securing full access to understanding general practitioners who are equipped with extra facilities, such as open referral channels to social services, may be helpful.
Although the socioeconomic inequalities in health in the Netherlands are small compared with those in many other countries, the actual inequalities are still great. Socioeconomic inequalities in health should be given a high priority in public health policy, and one of the merits of the research programme is that it forged a broad consensus on the need to develop specific policy measures.
This broad consensus may be surprising to those who live in countries such as the United Kingdom, where the debate on socioeconomic inequalities in health has become heavily politicised. The frosty reception of the Black report in the United Kingdom by the Conservative secretary of state to whom it was issued, contrasts with the reception of the results of the Dutch research programme. The committee overseeing the Dutch research programme was chaired by Professor L Ginjaar, chairman of the Health Council and former chairman of the Dutch Liberal party (the equivalent of the British Conservatives). He offered the results of the programme to the social democrat deputy minister, Hans Simons, who announced that he completely agreed with the programme committee's recommendations and that he would install a new programme committee chaired by Professor W Albeda, a former minister of social affairs belonging to the Christian Democrat party.
This lack of polarisation is difficult to explain, and is probably partly due to the Dutch political system, which can function only when coalitions are formed, and to the relative lack of class consciousness in the Netherlands. On the other hand, it is certainly also partly due to the deliberate efforts of those involved to avoid party political one sidedness.6 This tendency may seem to have resulted in an emphasis on the politically neutral lifestyle factors rather than the more politically sensitive material aspects of low socioeconomic status. In fact this difference in emphasis was due to a concern with finding the specific causal pathways leading from low socioeconomic status to ill health. Pathways including lifestyle factors are much easier to document than other causes, both because of their greater ease of measurement and because of their well established causal relation with ill health. The results of the research programme should not be interpreted as evidence for a greater importance of cultural or behavioural factors than structural or material factors1; differences in lifestyle may have their roots, at least partially, in differences in living conditions.
Egalitarian values are widespread in the Netherlands, especially with regard to health and disease, and those doing research and setting policy on socioeconomic inequalities in health generally believe that this basis for further action should be guarded preciously. The future will tell whether this consensus based approach is effective in reducing inequalities in health.
Most of the research reported in this paper was sponsored by the ministry of welfare, public health, and cultural affairs. I thank the research programme committee, chaired by Professor L Ginjaar, and especially its secretary Professor Paul van der Maas, who played a major part in getting the programme started and keeping it on track. I also thank Louise Gunning-Schepers, Dike van de Mheen, and Joost van der Meer for their useful comments and Karien Stronks for the data in table I. All views expressed in this paper are those of the author and do not necessarily reflect those of the programme committee or the ministry of welfare, public health, and cultural affairs.