Table 3.
Results of the binary logistic regression analysis (n = 247 patients).
| 95% CI for odds ratio | |||||||
|---|---|---|---|---|---|---|---|
| Variables | B | S.E. | Wald | z-value | Sig. | Odds ratio | Lower to Upper |
| Age (years) (Range = 21 to 94 years old) | 0.073 [0.02, 0.14] | 0.022 | 10.6 | 3.25 | 0.001 | 1.08 | 1.03 to 1.12 |
| Is mechanical ventilation needed? (1 = yes, 0 = no) | 5.26 [4.39, 7.82] | 0.77 | 46.4 | 6.81 | <0.001 | 193 | 43 to 878 |
| Pulse oximetry saturation at admission (%) (Range = 20 to 100%) | −0.043 [−0.10, −0.01] | 0.022 | 4.4 | −2.01 | 0.044 | 0.96 | 0.92 to 0.99 |
| The logarithm of dNLR at admission Log10(X + 1) [Range of Log10 (X+1)= 0.18 to 1.47] | 2.65 [0.46, 6.70] | 1.27 | 4.1 | −2.09 | 0.037 | 14.1 | 1.2 to 169.5 |
| The logarithm of platelets counts at admission Log10(X + 1) [Range of Log10 (X+1)= 1.46 to 3.26 x 109 cells per L] | −6.41 [−11.70, −3.28] | 2.02 | 10.0 | −3.17 | 0.002 | 0.002 | 0.00003 to 0.09 |
| Constant | 9.53 [0.67, 19.76] | 4.86 | 3.85 | 1.96 | 0.050 |
Each independent variable had 1 degree of freedom, including the constant. Brackets represent the 95% bootstrapped CI for the coefficients. Bootstrapped results are based on 1,000 bootstrapped samples. Three outliers (1%) were removed (standardized residuals ≥ 2.5). The variance inflation factor was the following for each predictor: age = 1.37; pulse oximetry saturation at admission = 1.05; mechanical ventilation = 1.80; logarithm of the platelet counts = 1.51; logarithm of dNLR = 1.16. Bayesian Information Criterion (BIC) = 132, (Hosmer & Lemeshow Goodness-of-Fit = χ2 = 6.11, df = 8, p = 0.64), j – 2 log likelihood = 99.0, Nagelkerke’s pseudo R2 = 0.83. In the first column, the minimum and maximum values in the dataset are presented within parentheses. So the model is accurate between each of these ranges. The [Brackets] in the second column represent the 95% CI.