Socioeconomic inequalities in morbidity and mortality in western Europe
1997; Elsevier BV; Volume: 350; Issue: 9076 Linguagem: Inglês
10.1016/s0140-6736(97)26033-2
ISSN1474-547X
AutoresDenny Vågerö, Robert S. Erikson,
Tópico(s)Health disparities and outcomes
ResumoJohan Mackenbach and colleagues' (June 7, p 1655)1Mackenbach J Kunst AE Cavelaars AEJM Groenhof F Geurts JJM EUWorking Group on Socioeconomic Inequalities in HealthSocioeconomic inequalities in morbidity and mortality in Western Europe.Lancet. 1997; 349: 1655-1659Summary Full Text Full Text PDF PubMed Scopus (910) Google Scholar conclusion that “Sweden and Norway have larger relative inequalities in health than most other countries” may be unjustified. One of us (DV) was a member of the international steering group for this study, but withdrew from the authorship; the other (RE) was indirectly involved because the social class classification used was intended to be an application of his and Goldthorpe's work.2Erikson R Goldthorpe J The constant flux. A study of class mobility in industrial countries. Clarendon Press, Oxford1992Google Scholar Comparisons of the ratio of manual/non-manual rates (relative risks) dominate mortality analyses in this study and are very important for its morbidity analyses; such comparison gives rise to several unresolved difficulties. We concentrate on the choice of indicators to measure health inequalities (relative rather than absolute). Figure 2 in Mackenbach's paper which is the basis for all general statements on the ranking of European countries, is based entirely on relative measures.First, all published estimates are prone to systematic error,3Kunst A Cavelaars A Groenhof F Geurts J Mackenbach J EU Working Group on Socioeconomic Inequalities in HealthSocioeconomic Inequalities in Morbidity and Mortality in Europe:a comparative study.Volume I:Main Report, Erasmus University Rotterdam. 1996; 35: 87-89Google Scholar but in this paper, the uncertainty margins refer only to chance variations. However, for ten of 11 countries the systematic error was in fact larger. If the margin of such error had been superimposed on the interval of random variation it would have been very difficult to draw any conclusions. Second, interpretation of relative risks without the corresponding absolute levels of mortality is very difficult. We present estimates of absolute levels of mortality for manual and non-manual workers, obtained from standardised mortality ratios for manual and non-manual workers3Kunst A Cavelaars A Groenhof F Geurts J Mackenbach J EU Working Group on Socioeconomic Inequalities in HealthSocioeconomic Inequalities in Morbidity and Mortality in Europe:a comparative study.Volume I:Main Report, Erasmus University Rotterdam. 1996; 35: 87-89Google Scholar and by combining them with number of deaths and persons at risk by age groups.4United Nations. Demographic Year Books, New York1984–92Google Scholar The figure corresponds to table 3 in Mackenbach's paper. Mortality is lowest for Swedish manual workers, and the manual/non-manual workers absolute risk difference is smallest in Norway. Danish, British, Italian, Swiss, and Spanish workers all have higher death rates than the corresponding groups in Sweden, but they are all more equal than in Sweden, according to Mackenbach's table 3.Mackenbach and colleagues do not discuss why the relative measure is the most relevant one, or whether it can be interpreted meaningfully in isolation. Rothman et al5Rothman K Greenland S Walker A Concepts of interaction.Am JEpidemiol. 1980; 112: 467-470PubMed Google Scholar discuss how different perspectives on interaction (here between class and country) may call for different methods of measurement; a perspective on public health prevention, rather than aetiological understanding, may call for focus on absolute levels and differences. It may be equally relevant that Swedish middle-aged male manual workers carry about 200 extra deaths per 100000 person years than non-manual workers, than that this represents a 1·4 relative risk.Inequality measures may be misleading if used to evaluate the effects of public policy, especially if they are presented in isolation. Mackenbach and colleagues say that “our data do not support the hypothesis that inequalities in health are smaller in countries whose social, economic, and health care policies are more influenced by egalitarian principles, such as Sweden and Norway” and “our results challenge widely held views on the relations between societal characteristics and the size of inequalities in health”. Whether or not these statements are true are of questionable relevance for an evaluation of policies or for a judgment of the effects of social inequality on health. The reason is that the large inequalities in Sweden and Norway in the 1980s largely seem to result from rapidly falling mortality rates among the upper non-manual classes rather than of high rates among the manual classes (data not shown).Nordic social policies have largely been based on the idea of equal treatment. Seemingly all groups have benefited, enjoying low death rates when compared with other west Europeans. We suggest that the potential benefit to manual classes of eradicating health inequalities is greater in countries such as Britain, Denmark, and Switzerland than in Norway and Sweden, despite the alleged low position of the last two in the west European league of health equality. Johan Mackenbach and colleagues' (June 7, p 1655)1Mackenbach J Kunst AE Cavelaars AEJM Groenhof F Geurts JJM EUWorking Group on Socioeconomic Inequalities in HealthSocioeconomic inequalities in morbidity and mortality in Western Europe.Lancet. 1997; 349: 1655-1659Summary Full Text Full Text PDF PubMed Scopus (910) Google Scholar conclusion that “Sweden and Norway have larger relative inequalities in health than most other countries” may be unjustified. One of us (DV) was a member of the international steering group for this study, but withdrew from the authorship; the other (RE) was indirectly involved because the social class classification used was intended to be an application of his and Goldthorpe's work.2Erikson R Goldthorpe J The constant flux. A study of class mobility in industrial countries. Clarendon Press, Oxford1992Google Scholar Comparisons of the ratio of manual/non-manual rates (relative risks) dominate mortality analyses in this study and are very important for its morbidity analyses; such comparison gives rise to several unresolved difficulties. We concentrate on the choice of indicators to measure health inequalities (relative rather than absolute). Figure 2 in Mackenbach's paper which is the basis for all general statements on the ranking of European countries, is based entirely on relative measures. First, all published estimates are prone to systematic error,3Kunst A Cavelaars A Groenhof F Geurts J Mackenbach J EU Working Group on Socioeconomic Inequalities in HealthSocioeconomic Inequalities in Morbidity and Mortality in Europe:a comparative study.Volume I:Main Report, Erasmus University Rotterdam. 1996; 35: 87-89Google Scholar but in this paper, the uncertainty margins refer only to chance variations. However, for ten of 11 countries the systematic error was in fact larger. If the margin of such error had been superimposed on the interval of random variation it would have been very difficult to draw any conclusions. Second, interpretation of relative risks without the corresponding absolute levels of mortality is very difficult. We present estimates of absolute levels of mortality for manual and non-manual workers, obtained from standardised mortality ratios for manual and non-manual workers3Kunst A Cavelaars A Groenhof F Geurts J Mackenbach J EU Working Group on Socioeconomic Inequalities in HealthSocioeconomic Inequalities in Morbidity and Mortality in Europe:a comparative study.Volume I:Main Report, Erasmus University Rotterdam. 1996; 35: 87-89Google Scholar and by combining them with number of deaths and persons at risk by age groups.4United Nations. Demographic Year Books, New York1984–92Google Scholar The figure corresponds to table 3 in Mackenbach's paper. Mortality is lowest for Swedish manual workers, and the manual/non-manual workers absolute risk difference is smallest in Norway. Danish, British, Italian, Swiss, and Spanish workers all have higher death rates than the corresponding groups in Sweden, but they are all more equal than in Sweden, according to Mackenbach's table 3. Mackenbach and colleagues do not discuss why the relative measure is the most relevant one, or whether it can be interpreted meaningfully in isolation. Rothman et al5Rothman K Greenland S Walker A Concepts of interaction.Am JEpidemiol. 1980; 112: 467-470PubMed Google Scholar discuss how different perspectives on interaction (here between class and country) may call for different methods of measurement; a perspective on public health prevention, rather than aetiological understanding, may call for focus on absolute levels and differences. It may be equally relevant that Swedish middle-aged male manual workers carry about 200 extra deaths per 100000 person years than non-manual workers, than that this represents a 1·4 relative risk. Inequality measures may be misleading if used to evaluate the effects of public policy, especially if they are presented in isolation. Mackenbach and colleagues say that “our data do not support the hypothesis that inequalities in health are smaller in countries whose social, economic, and health care policies are more influenced by egalitarian principles, such as Sweden and Norway” and “our results challenge widely held views on the relations between societal characteristics and the size of inequalities in health”. Whether or not these statements are true are of questionable relevance for an evaluation of policies or for a judgment of the effects of social inequality on health. The reason is that the large inequalities in Sweden and Norway in the 1980s largely seem to result from rapidly falling mortality rates among the upper non-manual classes rather than of high rates among the manual classes (data not shown). Nordic social policies have largely been based on the idea of equal treatment. Seemingly all groups have benefited, enjoying low death rates when compared with other west Europeans. We suggest that the potential benefit to manual classes of eradicating health inequalities is greater in countries such as Britain, Denmark, and Switzerland than in Norway and Sweden, despite the alleged low position of the last two in the west European league of health equality. Socioeconomic inequalities in morbidity and mortality in western EuropeAuthor's reply Full-Text PDF
Referência(s)