Final multivariable regression model after stepwise AIC reduction
| Analysis of maximum likelihood estimates | |||||
| Parameter | Estimate | SE | Wald Χ2 | Pr>Χ2 | |
| Intercept | 1 | 92.286697 | 235.9925 | 0.6859 | 0.4076 |
| Female sex | 1 | 0.5398305 | 1.489291 | 2.3955 | 0.1217 |
| Age | 1 | 0.9602131 | 1.053376 | 0.6092 | 0.4351 |
| BMI | 1 | 0.917319 | 1.10683 | 0.722 | 0.3955 |
| Age×BMI | 1 | 1.0025633 | 1.001641 | 2.4406 | 0.1182 |
| Admit SpO2 | 1 | 0.9231163 | 1.044982 | 3.3067 | 0.069 |
| Admit NLR | 1 | 1.1506189 | 1.056224 | 6.5748 | 0.0103 |
| Admit D-dimer | 1 | 0.9823575 | 1.013389 | 1.7929 | 0.1806 |
| Admit LDH | 1 | 2.3615071 | 1.331625 | 9.0006 | 0.0027 |
| Admit CRP | 1 | 2.0990818 | 1.476538 | 3.6211 | 0.057 |
| qSOFA | 1 | 2.261436 | 1.395566 | 5.9946 | 0.0143 |
LDH was categorised as normal, 1×<×<2× upper limit of normal (ULN), and >2× ULN. 85% concordance statistic was reached.
AIC, Akaike Information Criteria; BMI, body mass index; CRP, C reactive protein; LDH, lactate dehydrogenase; NLR, neutrophil to lymphocyte ratio; qSOFA, Quick Sequential Organ Failure Assessment; SpO2, oxygen saturation.