Welfare for all--or only for the needy?

Over a decade ago, an editorial in The Lancet voiced serious concern over the degree to which epidemiology had abandoned its traditional emphasis on issues of obvious importance to public health. Journal papers were increasingly occupied with refi ning statistical methods to address individual risk factors on a biological and molecular level. Appealing as the potential of molecular epidemiology might be, the editorial argued, “the benefi ts have not been, and are unlikely to be, at the population level”. The editorial called for a conscious eff ort to restore public health to epidemiology—by, for example, “reorienting its focus to global issues such as war, poverty, and environmental warming and to the social aspects of health and disease”. Multidisciplinary cooperation and between-population studies were among a range of suggestions. If this concern was a call for research with a broader scope and greater public-health relevance, the paper by Olle Lundberg and colleagues from the NEWS Nordic Expert Group in today’s Lancet could be a long-awaited answer, although it comes not from epidemiologists but from social scientists. Nonetheless, their extensive report to WHO’s Commission on Social Determinants of Health borders on social epidemiology and undoubtedly addresses a major public-health issue: is there any evidence to show that measures of health are related to welfare-state policies that are based on universal coverage, as in the Nordic countries? This question is not minor, because the alternative, targeted welfare for the needy, has been in vogue for some time, even on the Nordic scene. Lundberg analysed selected cross-national data from 18 countries of the Organisation for Economic Co-operation and Development (OECD) for infant and old-age mortality on an aggregate level to test the hypothesis that the design of welfare-state programmes and their level of generosity might aff ect these indices of population health. The original NEWS report (The Nordic Experience: Welfare States and public health) is long, technically complicated, and might not attract a wide readership. See Articles page 1633 In the short version presented today, the authors have selected a few core questions, described their methods so that critical readers can challenge the results, and provided an interpretation which is bound to raise further debate: they show that universal coverage and increased generosity in family policies are associated with lower rates of infant mortality, and generosity in basic universal pensions is associated with lower excess mortality in old age when both are compared with targeted welfare for needy people. Readers may fi nd it diffi cult to examine all the authors’ decisions about sampling, comparisons, and adjustments for confounders. Furthermore, the diff erences in mortality rates among most OECD countries are no longer striking. Those in southern Europe with higher mortality rates during the post-war years have largely caught up with northern European countries, and Spanish and Italian women now top the world league of lifetime expectancy. Welfare-state regimes are obviously only one of many conditions aff ecting public health and longevity. But, in Lundberg and colleagues’ paper, there was an association in favour of universal coverage, beyond the assumption that low levels of welfare generosity do aff ect mortality in outlier countries, such as New Zealand and the USA. Lundberg and colleagues’ paper does not explicitly address social inequalities in health and mortality within countries. The original report does, and shows that absolute levels of mortality in manual workers in Norway and Sweden are lower than in most other countries. However, a much debated issue is that there