No abstract

SUMMARY The association of weight for height and triceps skinfold with seven respiratory symptoms has been examined using logistic regression analysis in 7800 5 to 11 year old children (6200 in England and 1600 in Scotland). The (SDS).10 " The SDS is calculated for each child as the difference between his or her measurement and the average measurement of a population of the same age, sex, and country divided by the standard deviation of the measurement for that population. Here overweight and underweight refer to the two extremes of the range for weight for height, and fat and lean for triceps skinfold. The studied respiratory symptoms are shown in Table 1.Method of analysis. A logistic regression analysis was conducted for each respiratory symptom as a dichotomous dependent variable; the children being divided into two groups, those who had a positive response for the respiratory symptom and those who had a negative response. Children with a missing value were excluded. The analyses were conducted separately for weight for height and for triceps skinfold, each as a continuous independent variable. Many other independent variables were included in the analysis for each respiratory symptom because they had a possible association with respiratory symptoms. A sequence of linear models was fitted to the data to determine the extent to which the dependent variable was associated with each independent factor after adjustment for all other independent variables. The association with each obesity measure was also examined when adjustment was only made in the model for two of the independent variables, age and sex. This age-sex model was also

Among the harmful effects postulated for passive smoking is a possible association between parental smoking and respiratory conditions in children, which has been investigated in a large number of studies. A review article' concluded that the studies were consistent in suggesting increased infections in children under 1 year of age but inconsistent in older children. As almost all found some effect of parental smoking, the latter conclusion seems to have been due to the lack of a significant dose-response relation in just over half the studies considered.The studies on older children have varied in the symptoms studied, in the age range of the children, in the proportion of parents who smoked, and in the potentially confounding variables that have been taken into account. A report of a workshop on the effect of passive smoking on children2 listed nine groups of such variables that it is desirable to take into account. No study has included all of these, and most included only a few variables in one or two ofthe listed groups. This can be attributed largely to the fact that few2 of the studies were designed to investigate passive smoking effects, and were opportune analyses on data collected mainly to investigate the relations in children between symptoms and lung function and a variety of environmental factors.Of even more importance to the detection of a dose-response relation the studies have differed markedly in size and in the measure of passive smoking. The most usual measure was the number of parents smoking, providing lower power to detect a dose-response relation than a measure of the amount smoked. A recent review3 reported only three studies of young children in which the measure of passive smoking was cigarettes smoked per day, and just one study of older children.The National Study of Health and Growth, an on-going surveillance study of the health and growth of primary school children in England and Scotland, was also not designed to investigate passive smoking effects. Data on the number ofsmokers offive or more cigarettes a day in the child's home were collected in 1977 as a confounding variable in a study of the relation of respiratory illness and outdoor air pollution.4 These data also suggested a negative relation of child's height to number of smokers in the home after adjusting for birthweight.5 In order to study this association further, data on the number of cigarettes smoked at home by each parent, and by the mother during pregnancy, were collected in 1982.6 No data on lung function were obtained.Further examination of the 1977 data on English and Scottish children showed a number of statistically significant positive associations of respiratory symptoms with the number of smokers. Given the reasonable sample size, the availability of data for a 105Protected by copyright.

Uptake of school meals is very sensitive to changes in welfare policy. Monitoring the impact of these changes on children's health and welfare in families with low resources continues to be an important activity.

All patients attending the Royal Marsden Hospital with urothelial tumours have their occupational histories taken. About 10% of histories suggest a possibility of past exposure to urinary carcinogens, and on examination 4% indicate definite exposures justifying claims for Prescribed Industrial Disease benefit.