The correlation between air [size=7.721pt]pollutant exposures was estimated by Spearman[size=7.721pt]’[size=7.721pt]s [size=7.721pt]correlation coef[size=7.721pt]fifi[size=7.721pt]cients. Conditional logistic regres[size=7.721pt]sion models were implemented to quantify the [size=7.721pt]exposure-response associations between short-term [size=7.721pt]PM[size=5.4686pt]2.5[size=7.721pt], PM[size=5.4686pt]10[size=7.721pt], SO[size=5.4686pt]2[size=7.721pt], NO[size=5.4686pt]2[size=7.721pt], CO, and O[size=5.4686pt]3 [size=7.721pt]exposures and MI [size=7.721pt]mortality. Using data on case days and their matched [size=7.721pt]control days, we [size=7.721pt]fi[size=7.721pt]rst included exposure to each [size=7.721pt]pollutant as a natural cubic spline function (degrees [size=7.721pt]of freedom [size=7.721pt]¼ [size=7.721pt]3) in the model to plot exposure [size=7.721pt]response curves and examine possible nonlinear as as[size=7.721pt]sociations using likelihood ratio tests.
[size=7.721pt]求问论坛里的各位大佬！R语言clogit函数加上ns函数是如何做出剂量反应关系曲线的？我之前只会用RMS包做，但是RMS里的剂量反应曲线不支持条件逻辑回归。我用ns直接转换了自变量代入到clogit 的方程里，并不能得出相应的OR值啊。。。求问求问，可以有偿，谢谢各位！