Table 2 Independent predictors of hospital length of stay, ICU admission and death.

From: Defining the correlates of lymphopenia and independent predictors of poor clinical outcome in adults hospitalized with COVID-19 in Australia

Measures of outcome

Clinical variables

Model coefficienta

p-value

Length of stay

Age

0.07

 < 0.01

Haemoglobin

− 0.05

 < 0.01

Steroids during stay

1.79

 < 0.01

C-reactive protein (log10)

1.31

0.01

Presence of chronic liver disease

− 3.86

0.02

Lactate dehydrogenase (log10)

5.31

0.02

ICU admission

C-reactive protein (log10)

1.41

 < 0.01

Steroids before admission (yes)

1.49

 < 0.01

Lactate dehydrogenase (log10)

3.60

 < 0.01

Age

− 0.02

 < 0.01

Diabetic (yes)

0.70

0.02

Death

C-reactive protein (log10)

2.04

 < 0.01

Age

0.06

 < 0.01

Diabetic (yes)

1.05

0.01

  1. Clinical variables identified in our multivariate model as independent predictors of length of stay in hospital, ICU admission and death. All included variables have significant p-values (p < 0.05).
  2. aNote that the model coefficient gives an estimate for how one unit of change in the clinical variable alters either (a) the prediction of the outcome measure (for Length of Stay) or (b) the odds of the outcome being true (ICU admission or death).
Back to article page