Estimates of Rate Ratios (RR), 95%CI and Variance Components (VC) Obtained in Models (2) and (3), as Detailed in S2 Appendix, Using Individual and Aggregate Level Variables in the EPIC Study; Linear Regression Calibration.

a Variance component estimateb Also adjusted for weight, height and educational statusc Energy from non-fat and non-alcohol sourcesd current and former smokerse Modelled as cosine(Latitude)ref = reference category.Estimates of Rate Ratios (RR), 95%CI and Variance Components (VC) Obtained in Models (2) and (3), as Detailed in S2 Appendix, Using Individual and Aggregate Level Variables in the EPIC Study; Linear Regression Calibration.

a 方差组分估计（Variance component estimate）
b 同时针对体重、身高与受教育程度进行校正
c 源自非脂肪与非酒精来源的能量
d 当前吸烟者与既往吸烟者
e 以纬度的余弦值作为模型拟合项
ref = 参考类别（reference category）
在EPIC研究（EPIC Study）中，借助个体与聚合水平变量，通过附录S2中详述的模型（2）与模型（3），得到率比（Rate Ratios, RR）、95%置信区间（95%CI）及方差组分（Variance Components, VC）的估计值；分析采用线性回归校准法。