Table 1 Optimal model for Bacteremia in neonates
Binary logistic regression analysis (significant model p < 0.0001, Nagelkerke Correlation Coefficient 66%; N = 61 patients with a complete data set) | |||
|---|---|---|---|
Ln (Odds Bacteremia (Y / N)) = − 25.459 + 0.752(Maximum Blood Glucose [mmol/L]) + 0.119(Maximum Heart Rate [bpm]) + 0.108(% Bands) + 0.071(Maximum Neutrophils [× 109/L]) | |||
Variables in Final Binary Logistic Regression Equation | |||
Independent Variable | Odds Ratio | 95% Confidence Interval | CART breakpoint for association with bacteremia when parent node is maximum blood glucose |
Maximum Blood Glucose (mmol/L) | 2.121 | 1.182–3.806 | > 6 |
Maximum Heart Rate (bpm) | 1.127 | 1.040–1.221 | > 186 |
% Bands | 1.114 | 0.574–2.160 | > 2.15 |
Maximum Neutrophils (× 109/L) | 1.073 | 0.932–1.236 | > 11.7 |