Table 5.
Area under the curve for various hematological parameters in severe and very severe versus moderate COVID-19 patients
| Variable | Area under curve | p value | 95% Confidence Interval | |
|---|---|---|---|---|
|
| ||||
| Lower Bound | Upper Bound | |||
| WBC | .662 | .000 | .610 | .713 |
| RDWSD | .535 | .203 | .483 | .587 |
| RDWCV | .549 | .070 | .495 | .603 |
| NLR | .676 | <0.0001 | .623 | .728 |
| PLR | .654 | <0.0001 | .600 | .707 |
| MLR | .654 | <0.0001 | .601 | .707 |
| NLR and RDWSD | .861 | <0.0001 | .825 | .897 |
| NLR and RDWCV | .871 | <0.0001 | .838 | .904 |
| NLR and PLR | .859 | <0.0001 | .823 | .896 |
| NLR and MLR | .858 | <0.0001 | .821 | .894 |
The severe and very severe cases were taken as primary outcome for calculating area under the curve (AUC) for parameters. A new variable was generated by combining two parameters along with age and sex and were then subjected to binary logistic regression model with severity as outcome. The new variable was generated in form of predicted probability from the model.