Table 2.
Logistic regression calculated propensity scores.
| Variable | Patient number | Ferritin level median (IQR) | P-value |
|---|---|---|---|
| Model 1 | 126 vs. 126 | 944.3 (471.9–1778.25) vs. 687.1 (402.75–1179.5) | 0.0050 |
| Model 2 | 126 vs. 126 | 944.3 (471.9–1778.25) vs. 687.1 (402.75–1179.5) | 0.020 |
| Model 3 | 104 vs. 104 | 831.9 (426.1–1431.75) vs. 687.1 (392.23–1158.5) | 0.050 |
| Model 4 | 102 vs. 102 | 886 (445.58–1559.5) vs. 687.1 (391.08–1142.75) | 0.033 |
| Model 5 | 95 vs. 95 | 814.5 (417.5–1416.0) vs. 724.5 (391.65–1140.5) | 0.25 |
Each parameter was included in the model (Model 1: age, sex; Model 2: age, sex, CRP, WBC; Model 3: age, sex, length of hospitalization, intubation (or not), and maximal oxygen dose, CRP, WBC; Model 4: age, sex, length of hospitalization, intubation (or not), and maximal oxygen dose, CRP, WBC, D-dimer, creatinine, sodium, potassium, albumin, RBC, platelets, procalcitonin; Model 5: age, sex, length of hospitalization, intubation (or not), and maximal oxygen dose, CRP, WBC, d-dimer, creatinine, sodium, potassium, albumin, RBC, platelets, procalcitonin, remdesivir, steroid therapy, antibody cocktail therapy). IQR interquartile ranges.