Gravelle H. How much of the relation between population mortality and unequal distribution of income is a statistical artefact? BMJ 1998;316(31 January 1998):382-5.

This paper and the author's reply, basically do not deny the relationship, but want to have more studies to take account the influences of lifestyle, consumption, and environment in addition to income. Individual data are needed. Such studies are in process at present and preliminary analysis confirms the hypothesis. From a policy perspective, major decisions that have been made were made on much less confirmatory evidence. 

Summary points

The absolute income hypothesis, which states that the higher an individual's income, the lower his or her risk of mortality, is supported by a considerable body of evidence

However, the relative income hypothesis — that the distributon of income in a society affects the individual's risk of mortality — is being used increasingly in empirical work.

Associations between unequal income distribution and population health may be a statistical artefact resulting from the use of aggregate rather than individual data -- an example of the "ecological fallacy"

Because studies using population data cannot distinguish between the absolute and relative income hypotheses, the effects of income redistribution policies on population health can only be judged from individual data, interpreted by models of behaviours that affect health

Please click here to read complete text of the article