Univariable and multivariable ordinal regression: association of alveolar infiltrates and cavities with number of bacteria in sputum smear.

Dependent ordinal variable sputum smear ( -2Log 52, Chi2 0.015, Goodness of fit 0.88, Pseudo R2 (Cox and snell 0.084 and Nagelkerke 0.091), test of parallel lines 0.021. Model 2 -2Log 27.5, Chi2 0.068, Goodness of fit 0.98, Pseudo R2 (Cox and snell 0.037 and Nagelkerke 0.040), test of parallel lines 0.971. Model 3 -2Log 53, Chi2 0,020, Goodness of fit 0,083, Pseudo R2 (Cox and snell 0,067 and Nagelkerke 0,041), test of parallel lines 0,240. Model 4 -2Log 158, Chi2 0.08, Goodness of fit 0,012, Pseudo R2 (Cox and snell 0.11 and Nagelkerke 0.12), test of parallel lines 0.022. Model 5 -2Log 100, Chi2 0.009, Goodness of fit 0,080, Pseudo R2 (Cox and snell 0.11 and Nagelkerke 0.12), test of parallel lines 0.075. Reference categories: Sputum smear: 3+, Alveolar infiltrates: No infiltration, Cavities: 2 cavities.Univariable and multivariable ordinal regression: association of alveolar infiltrates and cavities with number of bacteria in sputum smear.

以痰涂片为因变量的有序分类变量（模型1）：-2对数似然值（-2Log）52，卡方（Chi2）值0.015，拟合优度（Goodness of fit）0.88，伪决定系数（Pseudo R2）：考克斯-斯内尔（Cox and Snell）为0.084、内格尔克尔克（Nagelkerke）为0.091，平行性检验（test of parallel lines）P值0.021；模型2：-2对数似然值27.5，卡方值0.068，拟合优度0.98，伪决定系数：考克斯-斯内尔为0.037、内格尔克尔克为0.040，平行性检验P值0.971；模型3：-2对数似然值53，卡方值0.020，拟合优度0.083，伪决定系数：考克斯-斯内尔为0.067、内格尔克尔克为0.041，平行性检验P值0.240；模型4：-2对数似然值158，卡方值0.08，拟合优度0.012，伪决定系数：考克斯-斯内尔为0.11、内格尔克尔克为0.12，平行性检验P值0.022；模型5：-2对数似然值100，卡方值0.009，拟合优度0.080，伪决定系数：考克斯-斯内尔为0.11、内格尔克尔克为0.12，平行性检验P值0.075。参照组设置为：痰涂片：3+，肺泡浸润：无浸润，空洞：2个空洞。单变量与多变量有序回归（ordinal regression）分析：探讨肺泡浸润、空洞与痰涂片细菌计数之间的关联。