Table 5. Multivariate logistic regression models for identifying associations of clinical and laboratory variables with actual mortality in 1,400 critically ill calves with diarrhea.

Variable	Coefficient	± SE	OR	95% CI for OR	P-value
Clinical and laboratory model 1(n = 1,385)
Intercept	-2.082	0.098
Ileus/Abdominal emergency	3.761	0.550	42.98	14.62–126.36	< 0.001
CNS involvement	2.978	0.347	19.65	9.96–38.76	< 0.001
Orthopedic problems	2.037	0.271	7.67	4.51–13.03	< 0.001
Cachectic body condition	1.640	0.181	5.16	3.62–7.35	< 0.001
Venous blood pH < 6.85	1.594	0.292	4.93	2.78–8.73	< 0.001
Laboratory model2 (n = 1,380)
Intercept	-1.958	0.096
Sodium concentration ≥ 151 mmol/L	1.462	0.195	4.32	2.95–6.32	< 0.001
Glucose concentration < 3.2 mmol/L	1.282	0.178	3.60	2.54–5.11	< 0.001
GGT activity < 31 U/L	1.075	0.198	2.93	1.99–4.32	< 0.001
Thrombocyte count < 535 G/L	0.846	0.180	2.33	1.64–3.31	< 0.001

Entered predictors were identified by means of Classification Tree analysis.

¹Hosmer-Lemeshow goodness-of-fit χ² = 2.89, df = 1, P = 0.089.

²Hosmer-Lemeshow goodness-of-fit χ² = 6.98, df = 3, P = 0.073