Mortality among residents near cokeworks in Great Britain

Objectives-To investigate whether residents near cokeworks have a higher st andardised mortality than those further away, particularly from cardiovascu lar and respiratory causes, which may be associated with pollution from cok eworks. Method-Cross sectional small area study with routinely collected postcoded mortality data and small area census statistics. Populations within 7.5 km of 22 cokeworks in Great Britain, 1981-92. Expected numbers of deaths withi n 2 and 7.5 km of cokeworks, and in eight distance bands up to 7.5 km of co keworks, were calculated by indirect standardisation from national rates st ratified for age and sex and a small area deprivation index, and adjusted f or region. Age groups examined were all ages, 1-14, 15-64, 65-74, greater t han or equal to 75. Only the 1-14 and 15-44 age groups were examined for as thma mortality. Results-There was a 3% (95% confidence interval (95% CI) 1% to 4%) excess o f all deaths within 2 km of cokeworks, and a significant decline in mortali ty with distance from cokeworks. The excess of deaths within 2 km was sligh tly higher for females and elderly people, but excesses within 2 km and dec lines in risk with distance were significant for all adult age groups and b oth sexes. The size of the excess within 2 km was 5% (95%CI 3% to 7%) for c ardiovascular causes, 6% (95%CI 3% to 9%) for ischaemic heart disease, and 2% (95% CI -2% to 6%) for respiratory deaths, with significant declines in risk with distance for all these causes. There was a non-significant 15% (9 5%CI -1% to 101%) excess in asthma mortality in the 15-44 age group. There were no significant excesses in mortality among children but 95% CIs were w ide. Within 2 km of cokeworks, the estimated additional excess all cause mo rtality for all ages combined related to region and mainly to the greater d eprivation of the population over national levels was 12%. Conclusions-A small excess mortality near cokeworks as found in this study is plausible in the light of current evidence about the health impact of ai r pollution. However, in this study the effects of pollution from cokeworks , if any, are outweighed by the effects of deprivation on mortality near co keworks. It is not possible to confidently exclude socioeconomic confoundin g or biases resulting from inexact population estimation as explanations fo r the excess found.