Predictive model for acute respiratory distress syndrome events in ICU patients in China using machine learning algorithms: a secondary analysis of a cohort study

Table 2 Baseline characteristics and clinical/laboratory parameters in the training and testing cohorts

Variable	Training cohorts (n = 236)	Testing cohorts (n = 60)	P value
Sex (male)	167 (70.8%)	36 (60%)	0.121
Age (year)	65.23 ± 18.02	66.03 ± 18.68	0.761
Minimum respiratory rate	23.14 ± 4.75	23.62 ± 5.08	0.494
Maximum respiratory rate	32.18 ± 6.08	31.65 ± 6.25	0.551
Minimum haematocrit	29.81 ± 6.62	30.52 ± 6.72	0.461
Minimum systolic pressure	99.40 ± 20.47	91.82 ± 19.50	0.010*
Minimum MAP	75.33 ± 60.14	67.25 ± 15.10	0.303
Maximum heart rate	122.33 ± 21.22	126.18 ± 21.37	0.211
Minimum glucose	135.00 ± 92.24	124.03 ± 54.68	0.378
Minimum white blood cell count	11.90 ± 6.24	12.88 ± 7.29	0.295
Minimum heart rate	90.00 ± 17.57	93.88 ± 17.85	0.129
Minimum temperature	36.70 ± 0.88	36.66 ± 0.84	0.738
Minimum sodium	140.06 ± 6.35	139.87 ± 6.82	0.834
APACHE II	19.96 ± 5.81	20.78 ± 5.51	0.322
PH	7.38 ± 0.10	7.37 ± 0.10	0.283
Bacteraemia	4 (1.69%)	3 (5%)	0.150
Diabetes	42 (17.8%)	17 (28.3%)	0.073
Tobacco	85 (36.0%)	24 (40%)	0.653

The binary variables are described as counts and percentages and were evaluated by the Chi-squared test or Fisher’s exact test. Continuous variables of each group are presented as the mean ± SEM. Student’s t-test was used to compare the normally distributed continuous variables. *P < 0.05, ARDS compared with non-ARDS
MAP mean arterial pressure, APACHE II Acute Physiology and Chronic Health Evaluation II

ISSN: 1479-5876