Letters

Smoking is part of a causal chain

While smoking is undoubtedly harmful to babies, the magnitude of the risk is less clear. The close correlation between adverse socioeconomic circumstances and smoking and between risk of the sudden infant death syndrome and deprivation requires that the analysis should take careful account of potential confounding. The importance of the association between the syndrome and deprivation is suggested by the univariate odds ratios associated with social factors quoted in the full report of this study²--for example, receipt of family income supplement, 6.27; income of <£200/ week, 3.57; living in rented accommodation, 3.81; and <0.5 rooms per person, 31.3. To avoid residual confounding, the measure of socioeconomic status used for adjustment in multivariate analyses needs to split the population into relatively homogeneous bands. To adjust for low socioeconomic status the authors seem to use receipt of family income supplement, dividing the population into two heterogeneous groups.³ Blackburn and Graham have shown, however, that even among women in receipt of income support (a more homogeneous group) the risk of smoking during pregnancy is strongly related to the degree of deprivation.⁴ This raises the potential for appreciable residual confounding.

While acknowledging that social variables remain significant after adjustment, the authors have chosen to concentrate on behavioural variables on the grounds that social variables are "not amenable to change." Between 1979 and 1987 the proportion of British children in families with less than half the mean household income increased from 12% to 26%.⁵ This suggests that social factors are amenable to change over quite short periods.

Smoking is harmful, but to lay responsibility for deaths due to the sudden infant death syndrome at the door of smokers, as the authors do in their conclusion, is to ignore the social context in which people smoke. There is considerable evidence that both taking up and stopping smoking are strongly influenced by material circumstances and that exhortation is ineffective in reducing rates of smoking among the most vulnerable people. Although parental smoking is probably causally related to the sudden infant death syndrome, if we are to intervene effectively it is important to recognise that smoking is part of a causal chain in which social factors are the distal cause and to direct equal attention to these social causes rather than blame the victims.

Senior lecturer in paediatric epidemiology Department of Paediatric Epidemiology, Institute of Child Health, London WC1N 1EH

Professor of community child health Lecturer in applied social studies Department of Applied Social Studies, University of Warwick, Coventry CV4 7AL

Stuart Logan, Nick Spencer, Clare Blackburn 

1. Blair PS, Fleming PJ, Bensley D, Smith I, Bacon C, Taylor E, et al. Smoking and the sudden infant death syndrome: results from the 1993-5 case-control study for confidential inquiry into stillbirths and deaths in infancy. BMJ 1996;313:195-8. (27 July.) [Abstract/Free Full Text]
2. National Advisory Body for CESDI. Annual report for 1994. London: Department of Health, 1996.
3. Fleming PJ, Blair PS, Bacon C, Bensley D, Smith I, Taylor E, et al. Environment of infants during sleep and risk of the sudden infant death syndrome: results from the 1993-5 case-control study for confidential inquiry into stillbirths and deaths in infancy. BMJ 1996;313:191-5. (27 July.) [Abstract/Free Full Text]
4. Blackburn C, Graham H. Achieving lifestyle change: a preliminary report. Warwick: Department of Applied Social Studies, University of Warwick, 1995.
5. Oppenheim C. Poverty: the facts. London: Child Poverty Action Group, 1990.