Skip to main content

Table 3 Models for predicting each feature postVPS (after 6 months), given their value at preTT (baseline), or both their value at preTT and their change after the TT (postTT)

From: Gait apraxia evaluation in normal pressure hydrocephalus using inertial sensors. Clinical correlates, ventriculoperitoneal shunt outcomes, and tap-test predictive capacity

 

postVPS ~ preTT

postVPS ~ preTT + ΔpostTT-preTT

Likelihood ratio test

 

R2adj

R2adj

χ2

p-value

Clinical

    

 nSteps2turn

0.13

0.14

1.409

0.235

 Tinetti Balance

0.27

0.30

2.661

0.103

 Tinetti Gait

0.18

0.17

0.529

0.467

 Tinetti Total

0.33

0.41

6.213

0.013

 GSS

0.23

0.23

0.637

0.425

 iNPH-GS

0.16

0.14

0.083

0.773

TUG

 TestDuration

0.11

0.38

16.437

0.0001

 TotalSteps

0.29

0.29

0.914

0.339

 WalkTime

0.13

0.42

18.151

0.00002

 StandTime

-0.02

0.18

10.333

0.001

 SitTime

0.00

0.20

10.392

0.001

 TurnSteps

0.44

0.43

0.50

0.480

 Cadence

0.26

0.39

8.714

0.003

 StrideLength

0.48

0.53

5.715

0.017

 DoubleSupport

0.35

0.35

1.386

0.239

 GaitSpeed

0.28

0.39

8.163

0.004

 TrunkInclination

0.30

0.49

14.693

0.0001

 maxTC1

0.25

0.23

0.150

0.699

 maxTC2

0.18

0.32

8.996

0.003

 minTC

0.17

0.32

9.815

0.002

 PitchAtTC2

-0.02

0.08

5.465

0.019

 pci

0.05

0.19

7.338

0.007

 StrideSD

-0.02

0.00

2.016

0.156

 psdF

0.48

0.52

3.929

0.047

 psdW

0.20

0.28

5.914

0.015

18mW

 TestDuration

0.37

0.37

1.214

0.271

 TotalSteps

0.52

0.51

0.369

0.544

 Cadence

0.38

0.49

8.882

0.003

 StrideLength

0.39

0.39

0.792

0.374

 DoubleSupport

0.32

0.35

2.761

0.097

 GaitSpeed

0.32

0.35

2.761

0.097

 TrunkInclination

0.56

0.70

16.055

0.0001

 maxTC1

0.27

0.26

0.811

0.368

 maxTC2

0.28

0.26

0.016

0.898

 minTC

0.31

0.30

0.667

0.414

 PitchAtTC2

− 0.03

− 0.05

0.153

0.695

 pci

0.05

0.12

4.311

0.038

 StrideSD

0.12

0.11

0.281

0.596

 psdF

0.27

0.42

10.174

0.001

 psdW

0.27

0.43

10.901

0.001

  1. R2adj indicates the model goodness of fit. Likelihood ratio tests between nested models indicate the added prognostic value of CSF-TT information. Significant p-values (< 0.05) are highlighted in bold