论文部分内容阅读
目的应用地理加权回归模型探讨气象因素和大气污染因素影响我国女性肺癌发病的空间属性。方法从《2012中国肿瘤登记年报》获取2009年我国72个肿瘤登记点的肺癌发病数据,分别从中国气象局和《2008中国环境年鉴》获取2008年气象和大气污染资料。使用因子分析将收集的气象、大气污染等空间因素降维,消除多重共线性。以标准化发病率作为因变量、城乡类型及空间因素因子得分作为自变量,分别拟合泊松回归模型和地理加权回归模型。结果东北地区尤其是辽东半岛为我国女性肺癌的高发区域,具有明显的地区集聚性。城乡类型、纬度指向因子(PM10、温度、降水量)与女性肺癌发病的关联具有统计学意义,农村相对于城市为保护因素(RR=0.758,P=0.003);纬度指向因子为危险因素,因子得分越高(PM10越高,温度越低,降水量越小),女性患肺癌风险越大(RR=1.104,P=0.027)。地理加权回归模型结果进一步显示,城乡类型对我国女性肺癌发病的影响强度存在东西地区差异,纬度指向因子影响强度存在南北地区差异。结论地理加权回归模型对具有空间自相关性的数据获得更优的拟合,可以较好地揭示空间因素在地区间作用的差异。
Objective To explore the spatial attributes of the incidence of lung cancer in women in China using the geographic weighted regression model. Methods The data of lung cancer incidence in 72 tumor registration sites in China were obtained from “2012 China Cancer Register Annual Report”, and the 2008 meteorological and atmospheric pollution data were obtained from China Meteorological Administration and “2008 China Environment Yearbook” respectively. Using factor analysis, the spatial factors such as meteorology and air pollution collected will be reduced to eliminate multicollinearity. Standardized incidence was taken as dependent variable, urban-rural type and spatial factor score as independent variable to fit Poisson regression model and geographic weighted regression model respectively. Results Northeast China, especially the Liaodong Peninsula, is a high incidence area of female lung cancer in China, with obvious regional agglomeration. The correlation between urban-rural type, latitude-oriented factor (PM10, temperature, precipitation) and the incidence of lung cancer in females was statistically significant, while that in rural areas was protected relative to that in urban areas (RR = 0.758, P = 0.003); latitude-oriented factors were risk factors, The higher the score (higher PM10, lower temperature, less precipitation), the greater the risk of lung cancer in women (RR = 1.104, P = 0.027). Geographic weighted regression model results further show that the impact of urban and rural types on the incidence of lung cancer in our country there are differences between east and west, latitude factor of influence there is the difference between north and south. Conclusion Geographic weighted regression model can obtain a better fitting result for the data with spatial autocorrelation, which can better reveal the difference of spatial factors in different regions.