Machine learning approach for predicting post-intubation hemodynamic instability (PIHI) index values: towards enhanced perioperative anesthesia quality and safety

Table 2 The statistics of the input features in the training and testing sets

Features	Training sets	Testing sets	P value ^a
Sex
Male (1)	30.17%	29.97%	0.748
Female (0)	69.83%	70.03%
Age	49.98 (13.88)	50.07 (13.96)	0.611
Height (cm)	164.20 (6.72)	164.22 (6.75)	0.882
Weight (kg)	64.51 (11.32)	64.52 (11.33)	0.922
BMI	23.84 (3.38)	23.84 (3.32)	0.907
Hypertension
Yes (1)	17.01%	17.64%	0.206
No (0)	82.99%	82.36%
Pre SP (mmHg)	133.55 (20.41)	133.26 (20.51)	0.282
Pre DP (mmHg)	78.42 (12.12)	78.49 (12.16)	0.624
Pre HR (BPM)	79.06 (13.08)	79.03 (13.12)	0.874
Fentanyl (μg/kg)	1.76 (0.53)	1.77 (0.54)	0.494
Lidocaine (mg/kg)	0.578 (0.248)	0.578 (0.249)	0.894
Propofol (mg/kg)	2.23 (0.45)	2.22 (0.44)	0.732
Rocuronium (mg/kg)	0.730 (0.113)	0.728 (0.113)	0.333

The sex and hypertension are discretized to 0–1 value and presented in percentage; Other continuous features are described by mean value (standard deviation, Std); The training and testing sets exhibited a comparable distribution, with no significant differences according to T-test result.
Pre preoperative, SP systolic pressure, DP diastolic pressure, HR heart rate, BMI body mass index, BPM beats per minute.
^aP value of T-test between the training and testing sets

ISSN: 1471-2253