The evidence here, as elsewhere, suggests that education is certainly relevant, but more because better education is associated with general differences in patterns of life than because discrete parts of a lifestyle can be changed. Health-change policies which focus entirely on the individual may be ineffective not only because exposure to health risks is largely involuntary, but also, as this study has shown, because of unwarranted assumptions about the extent to which behaviour can, in these circumstances, be effective in improving health.