Spatial and Temporal Analysis of Lung Cancer Mortality in Xuan Wei, China

Among all types of cancer, lung cancer deserves special attention because it causes nearly one third of the total cancer deaths worldwide. In the current study area of Xuan Wei County, China, lung cancer mortality rate was among the highest in China for both sexes according to the first national mortality survey in 1973-1975, which was followed by a large scale primary prevention measurement, stove improvement. Local residents changed from unvented open firepits to stoves with chimneys in order to reduce the indoor air pollution caused by bituminous coal burning, because it has been linked to the high lung cancer risk in this area. In terms of the effectiveness of the intervention, there was some discrepancy between previous findings. One cohort study in central Xuan Wei with highest lung cancer rates suggested a lung cancer risk reduction due to the stove improvement; while the county-wide mortality surveys indicated an apparently increasing lung cancer rates. This discrepancy could result from some complex spatio-temporal changes in lung cancer rates in Xuan Wei in the past decades. Theoretically, if the stove improvement was overall effective, the spatial pattern should have disappeared or changed. Systematic analyses of the temporal and spatial pattern changes in lung cancer mortality in this area are warranted. This thesis examined the temporal trend of lung cancer mortality rates over the past three time periods of 1976-79, 1990-92 and 2004-05 by using age-period-cohort analysis with Poisson model. It also examined the spatial pattern changes at the village levels over the three time periods using Kulldorff's spatial scan statistics. The main findings are: 1) there was a decreasing trend in pathologically confirmed lung cancer mortality, and an increasing trend in chest X-ray diagnosed lung cancer mortality over the three periods; and 2) there was still a statistically significant spatial lung cancer mortality cluster in the central part of Xuan Wei, and it was consistent over the three periods, but the spatial variations became weaker with time. These findings have important implications for evaluating the effectiveness of the intervention measures taken to reduce lung cancer risk and formulating new etiological hypothesis of the lung cancer epidemic in Xuan Wei, China.