Table 2.

Estimated coefficients (Est) with their SE and p value

Predictor	Clinical data alone: day 1			Clinical data+day 1*, biomarker data			Day 3			Day 4			Day 5
	Est	SE	P value	Est	SE	P value	Est	SE	P value	Est	SE	P value	Est	SE	P value
Intercept	-5.37	1.46	0.0003	−4.36	1.35	0.001	0.67	1.97	0.73	0.005	1.86	0.99	−0.20	1.68	0.9
Log Neut/Lymp D1				0.28	0.16	0.08
Log Neut/Lymp D3							0.41	0.19	0.03
Log Neut/Lymp D4										0.48	0.2	0.02
Log Neut/Lymp D5													0.52	0.21	0.01
Log Urea/Creat D2							−4.22	1.24	0.0007
Log Urea/Creat D3							5.13	1.30	0.0001
Log Urea/Creat D4										1.08	0.35	0.002	−4.97	1.72	0.0003
Log Urea/Creat D5													6.32	1.77	0.0004
Age (years)	0.064	0.012	<0.0001	0.073	0.012	<0.0001	0.069	0.013	<0.0001	0.071	0.012	<0.0001	0.066	0.012	<0.0001
O2 saturation	−0.028	0.012	0.01	−0.03	0.012	0.01	−0.03	0.013	0.03	−0.03	0.012	0.02
Respiratory rate	0.085	0.022	0.0001	0.085	0.022	0.0001	0.08	0.023	0.0003	0.087	0.022	0.0001	0.09	0.022	0.0001
Smoking	0.7	0.263	0.0073	0.7	0.267	0.01	0.8	0.27	0.004	0.71	0.27	0.008	0.76	0.28	0.006

Note that different variables are selected at different days so that models are not nested.

* The addition of biomarker data on day 2 did not contribute any additional predictive power of that obtained on day 1.

Creat, creatinine; D, day; Neut/Lymp, neurotrophil/lymphocyte; O₂, oxygen.