Air pollution and other local factors in respiratory disease.

Abstract

The mortal risks of acute episodes such as the London smog of December, 1952, are well recog nized, but the insidious effects of lesser but more continuous atmospheric pollution are less clearly understood. Contrasts in bronchitic mortality between Great Britain and Scandinavia, and between town and country in the United Kingdom, strongly suggest that such effects are important. Uncovering the aetiology of a slowly progressive disease like chronic bronchitis involves the study of its evolution from trivial illness to ultimate death and its relation to other respiratory diseases. As a step towards this, we have used the sickness absence experience of British civil servants to supplement the usual mor tality data available from the Registrar General's publications. The British Civil Service offers considerable advantages as a population for epidemiological study. Numbering some 600,000, it has standard conditions of sick pay and superannuation and con tains large groups, uniform in pay and job, widely distributed throughout the United Kingdom. The individual sickness absence records, giving the dates of onset and return to work and the certified diag nosis, are available for the whole of an employee's service and are afterwards retained for 10 years. They present a unique opportunity both for longi tudinal studies of the natural history of disease and for the more usual cross-sectional study of current morbidity experience. The potential value of these data is evident in Fig. 1 (opposite), which shows the distribution of the average time lost through sickness absence by postmen in different parts of the United Kingdom. This morbidity pattern again suggests the hazards to health of urban life which are so clearly implied by the urban-rural contrast in bron chitis mortality. The limitations of such material must, however, be emphasized. Any occupational group is selected both by the individual's choice of career and by the em ployer's policy of recruitment and discharge. In the Civil Service such policy may vary between different occupations and at different periods. The effect of changes in retirement policy on sickness rates in the Post Office from 1891 to 1946 has been shown by Roberts (1948). Moreover, some illnesses may in capacitate, whether temporarily or permanently, those in arduous occupations but not those doing light indoor work. Sickness rates should thus be used with caution as measures of relative occu pational risk or trends in morbidity. In ascribing importance to the different sickness rates of postmen and indoor workers, we have therefore looked for changes in this difference with age, locality, or diag nosis. No secular comparison of sick rates has been made. On the other hand, uniform medical standards of recruitment and ill-health retirement are applied by the Civil Service Commission and the Treasury Medical Service over the whole country, and com parisons by locality are largely free from bias due to staff selection as well as from that due to pay and job. In this study respiratory mortality in the middle aged population at large is first correlated with measures of air pollution, population density, and domestic overcrowding, in different areas of the United Kingdom. Wastage and sickness rates among civil servants are then related to the same indices in the same areas. There follows a more detailed analysis of morbidity among the same civil servants at different ages and in areas with con trasting degrees of air pollution.