Exploring the incidence of lung cancer in small areas across Scotland

Abstract

Lung cancer is one of the most important causes of 'avoidable deaths' globally and is responsible for approximately 900,000 deaths per year. However, lung cancer rates tend to be higher for males than for females and the disease also varies geographically, as rates are far higher in developed countries compared to developing countries. Scotland has the highest rate of lung cancer of any country where lung cancer data is available. However, explaining the spatial distribution of this disease is difficult because lung cancer has a number of known causes that operate at a range of different spatial scales. This is further complicated by the lag time between the period of exposure to a risk factor and the date of diagnosis. This thesis examines the causes of lung cancer across Scotland, using lung cancer registrations for the period 1988 to 1991. Exploratory methods are presented for examining the geographical distribution of the disease in small areas using methods of age-standardisation and cluster detection to identify areas with unusual rates. Estimates of the key risk factors potentially associated with lung cancer are calculated for the same small areas. These include estimates of smoking behaviour, air pollution levels in 1971 and 1991, radon gas potential, coal mining activity, quarrying activity and area deprivation. The risk factors are incorporated into a set of regression models to examine which factors are significant in explaining lung cancer incidence. Finally, the residual values derived from the optimum model of lung cancer incidence in Scotland are examined to identify areas where lung cancer incidence is particularly high and low. This study revealed that there were marked geographical differences in lung cancer rates, with higher rates in the large urban areas, especially Glasgow, and also the more deprived areas of Scotland. Smoking was consistently significant in explaining lung cancer incidence for all cohorts, types of lung cancer and urban-rural areas. The estimated air pollution levels in 1971 were also found to be significant, but the 1991 estimates were not. Exposure to radon was only significant in explaining lung cancer in the younger age groups. However, the coal mining and quarrying variables did not independently influence the incidence of the disease. Area deprivation and measure of urbanness both had a significant effect on lung cancer incidence in Scotland that was independent of the key risk factors. The analysis of the residual values showed that, having controlled for the key risk factors, lung cancer incidence is higher than expected in rural rather than urban areas.