Predictive determinants of overall survival among re-infected COVID-19 patients using the elastic-net regularized Cox proportional hazards model: a machine-learning algorithm

Table 3 The selected features of time to in-hospital death or discharge and the regularized elastic-net coefficients in the re-infected COVID-19 patients (α_optimal=0.9 and λ_optimal= 0.03985)

Features	Coefficient ^a
Type of patient transfer (by EMS)	0.9145
SpO₂ (≤85%)	0.8145
Intubation (yes)	0.5699
Triage level (level 1 vs. others)	0.5067
Creatinine (mg/dL)	0.3385
WBC count (×10⁹ cells/L)	0.0098
BUN (mg/dL)	0.0082
LDH (U/L)	0.0003

Abbreviations: BUN blood urea nitrogen, EMS Emergency Medical Services, LDH lactate dehydrogenase, SpO₂ saturation of peripheral oxygen, WBC white blood cell
^a Estimated coefficients using regularized elastic-net analysis sorted by magnitude from highest to lowest

ISSN: 1471-2458