论文部分内容阅读
目的探讨男性工人肺功能预计值方程的建立方法。方法选择浙江省内5家工厂303名男性工人建立肺功能预计值方程,对普通最小二乘回归(OLS回归)及稳健回归的结果进行比较,并对预计值方程的应用效果进行探讨。结果 OLS回归建立的预计值方程为:FVC=-6.024 0+0.064 0×height+0.003 3×weight-0.020 2×age,FEV1=-4.223 0+0.053 5×height-0.003 0×weight-0.027 3×age,FEV1/FVC%=102.868 6-0.143 7×weight-0.237 5×age,MMEF=-1.213 7+0.040 5×height-0.045 1×age,PEF=-8.364 2+0.098 5×height,FEF25=-7.819 8+0.090 8×height,FEF50=-1.804 8+0.046 8×height-0.040 0×age,FEF75=-0.078 6+0.020 5×height-0.038 9×age。稳健回归建立的预计值方程为:FVC=-6.150 0+0.064 3×height+0.003 8×weight-0.019 2×age,FEV1=-4.535 0+0.055 5×height-0.004 1×weigh-0.026 3×age,FEV1/FVC%=102.379 0-0.146 6×weight-0.217 3×age,MMEF=-1.799 5+0.042 7×height-0.040 9×age,PEF=-7.677 7+0.094 4×height,FEF25=-7.659 9+0.088 8×height,FEF50=-2.851 0+0.051 4×height-0.034 8×age,FEF75=0.086 0+0.017 9×height-0.034 7×age。稳健回归对FEF75和FEV1的拟合效果高于OLS回归。删除残差异常的观察值后,稳健回归的判定系数(R2)未发生变化,OLS回归的R2得到改善。结论稳健回归可建立可靠的肺功能预计值方程,具有一定的实用价值。
Objective To investigate the method of establishing the predictive value of lung function in male workers. Methods 303 male workers in 5 factories in Zhejiang Province were selected to establish the pulmonary function prediction equation. The results of ordinary least squares regression (OLS) regression and robust regression were compared and the effects of the prediction equations were discussed. Results The predicted value equation established by OLS regression was: FVC = -6.024 0 + 0.064 0 × height + 0.003 3 × weight-0.020 2 × age, FEV1 = -4.223 0 + 0.053 5 × height-0.003 0 × weight-0.027 3 × age, FEV1 / FVC% = 102.868 6-0.143 7 × weight-0.237 5 × age, MMEF = -1.213 7 + 0.040 5 × height-0.045 1 × age, PEF = -8.364 2 + 0.098 5 × height, FEF25 = 7.819 8 + 0.090 8 × height, FEF50 = -1.804 8 + 0.046 8 × height-0.040 0 × age, FEF75 = -0.078 6 + 0.020 5 × height-0.038 9 × age. The predicted regression equation established by robust regression is: FVC = -6.150 0 + 0.064 3 × height + 0.003 8 × weight-0.019 2 × age, FEV1 = -4.535 0 + 0.055 5 × height-0.004 1 × weigh-0.026 3 × age , FEV1 / FVC% = 102.379 0-0.146 6 × weight-0.217 3 × age, MMEF = -1.799 5 + 0.042 7 × height-0.040 9 × age, PEF = -7.677 7 + 0.094 4 × height, FEF25 = -7.659 9 + 0.088 8 × height, FEF50 = -2.851 0 + 0.051 4 × height-0.034 8 × age, FEF75 = 0.086 0 + 0.017 9 × height-0.034 7 × age. Robust regression fitted FEF75 and FEV1 better than OLS regression. After removing the observed abnormal residuals, the coefficient of determination (R2) of robust regression did not change, and the R2 of OLS regression was improved. Conclusions Steady regression can establish a reliable predictive equation of pulmonary function, which has some practical value.