Department of Social Science and Medicine, Imperial College London, St Dunstans Road, London W6 8RP, UK. E-mail: d.blane{at}imperial.ac.uk
Bopp and Minders paper1 is the most welcome first report from the Swiss National Cohort Study. This study links data from the 1990 Swiss national census with subsequent death records, so far 19901997. The study supplements the cross-sectional Swiss Health Survey, providing the first longitudinal data on the social distribution of health, as indexed by mortality, in Switzerland.
Sizeable gradients in mortality by education are demonstrated at ages 25 years to 75+ years. At ages 4054 years, for example, the standardized mortality ratio of male graduates is 59, compared with 159 for men who did not record any secondary education. Some abatement of this relationship is found among women, and at older ages. As Bopp and Minder detail meticulously, these findings are consistent with those from most other rich countries. They point to the relevance of these results for health policy and for the design of interventions to reduce social inequalities in health. My comments concern these hopes.
Minimizing the number of people receiving low levels of education is the most obvious policy response to Bopp and Minders evidence. As those receiving little education usually are the more disadvantaged members of the population, the appropriate policy initiatives will enhance the educational facilities in working class and immigrant communities. Initiatives might include free nursery education, lower teacher to pupil ratios, free school meals, and wages for those who stay in school after the minimum school leaving age.
The design of further policy initiatives will depend on a more detailed understanding of the educationhealth relationship. At least five causal processes could contribute to the regularly observed association between education and health.2 The material and cultural resources of the family of origin have a major influence on a childs educational attainment, so the association between education and health could be due to the long-term influence on adult health of childhood circumstances. Second, educational qualifications are a strong predictor of occupation and income during adulthood, so the association between education and health could be due to the contemporaneous influence on adult health of adult circumstances. Third, length of education may influence receptivity to health education messages, either because such messages are written in the language of the educated, or because the material and cultural resources of the educated allow them more easily to adapt their behaviour. Fourth, the association between education and health could be due to a third, background, variable which influences both the capacity to complete a prolonged period of formal education and the capacity to maintain health and cope with disease. Social psychological constructs, such as self-efficacy and time preference, have been suggested; although, presumably, affluence and enthusiasm for education in the family of origin would have a similar effect. Finally, ill health during childhood could both limit educational attainment and predispose towards adult morbidity and premature mortality. Each of these several processes is plausible biologically and socially, but each has quite different implications for health policy and the design of interventions to reduce inequalities in health. The policy and intervention hopes of Bopp and Minder, in consequence, should be shaped by attempts to unravel the causal processes which link education and health.
The most sustained investigation of these questions has been conducted by Professor Jerry Morris at the London School of Hygiene and Tropical Medicine. In an ecological study,2 based on local education authority areas, Morris and his colleagues demonstrated a strong inverse relationship between deprivation and examination results at age 16 years. Rank correlations were found of 0.93 and 0.89 between education and, respectively, the Carstairs and Townsend index; social deprivation thus explaining statistically more than three-quarters of the variance in educational attainment. The massive impact of deprivation on education is known to government,3 but is acknowledged rarely in policy debates. Education showed similar, if weaker, correlations with the various measures of mortality; ranging from 0.60 (infant mortality) to 0.77 (male all-cause mortality). In other words, the educational attainment of adolescents predicts the mortality rate of their parents and grandparents generation, as well as that of their younger siblings. In multiple regression analyses, controlling for deprivation weakened the educationmortality relationship more than controlling for education weakened the deprivationmortality relationship; suggesting that material circumstances are more important to mortality risk than education (a conclusion supported elsewhere4). Because the results of the multiple regression analyses varied somewhat, depending on whether the index of Carstairs or Townsend was used, the study was repeated using the Department of the Environments summary index of local conditions in England.5 Once again, deprivation not education had the stronger association with all-cause mortality; although an education effect remained for coronary heart disease mortality and infant mortality.
Morris followed these ecological analyses with a study using individual-level data.6 In these analyses, the association of self-reported health status with deprivation was stronger than with educational attainment. An independent effect of education remained, however, for longstanding illness among men, but not women, and for self-assessed health. In terms of the causal processes sketched above, Morris investigations show that a considerable proportion of the relationship between education and health can be accounted for by education acting as a marker of adult socioeconomic circumstances. Education retains an independent effect, however. Which of the other causal processes account for this independent effect remains an open question, although the ecological cross-generation finding points to the possible importance of childhood circumstances. Morris planned a more direct investigation of the influence of childhood circumstances, together with the influence of age at health behaviour change, potential third factors and childhood illness, in one of the British birth cohort studies. Support for these investigations, shamefully, was refused. As a result, the baton is lying there, waiting for someone else to pick up.
![]() |
References |
---|
![]() ![]() |
---|
2 Blane D, White I, Morris JN. Education, social circumstances and mortality. In: Blane D, Brunner E, Wilkinson R (eds). Health and Social Organisation. London: Routledge, 1996, pp. 17187.
3 Department of Education and Science. Statistical Bulletin 1984, 13/84.
4 Davey Smith G, Hart C, Hole D et al. Education and occupational social class: which is the more important indicator of mortality risk? J Epidemiol Community Health 1998;52:15360.[Abstract]
5 Morris JN, Blane D, White I. Levels of mortality, education and social conditions in the 107 local education authority areas of England. J Epidemiol Community Health 1996;50:1517.[Abstract]
6 White I, Blane D, Morris JN, Mourouga P. Educational attainment, deprivation-affluence and self-reported health: a cross-sectional study. J Epidemiol Community Health 1999;53:53541.[Abstract]