Table 3 Evaluation of different models for estimating BP from PAT or PTT.

From: Pulse arrival time as a surrogate of blood pressure

 

A posteriori models

Population-based models

Model

RMSE

MAE

MAD

Model

RMSE

MAE

MAD

PAT

BP = \(\frac{{\text{a}}}{{\text{PAT}}}\) + b

SBP

5.50

3.63

4.14

Poon16

SBP

11.04

7.10

8.46

DBP

4.49

3.42

2.92

DBP

8.37

6.12

5.71

BP = \(\frac{{\mathbf{a}}}{{\mathbf{PAT}}^2}\) + b

SBP

5.48

3.61

4.12

Gesche18

SBP

43.76

21.63

38.06

DBP

4.49

3.42

2.92

DBP

46.70

22.93

40.71

BP = a\(\times \)ln(PAT) + b

SBP

5.53

3.66

4.15

Fung37

SBP

8.51

5.70

6.33

DBP

4.50

3.42

2.92

DBP

8.02

5.90

5.43

PTT

BP = \(\frac{{\text{a}}}{{\text{PTT}}}\) + b

SBP

3.98

2.82

2.81

Poon16

SBP

8.49

5.40

6.56

DBP

4.02

3.12

2.54

DBP

7.72

5.64

5.26

BP = \(\frac{{\mathbf{a}}}{{\mathbf{PTT}}^2}\) + b

SBP

3.91

2.78

2.76

Gesche18

SBP

4542

1773

4185

DBP

4.01

3.12

2.53

DBP

4546

1776

4188

BP = a\(\times \)ln(PTT) + b

SBP

4.05

2.86

2.87

Fung37

SBP

30.69

20.77

22.61

DBP

4.03

3.12

2.55

DBP

33.80

22.45

25.29

  1. All values are given in units of mmHg. Best performing models for PAT and PTT are highlighted in bold.