Table 3 Six independent predictors of two clusters identified using the K-means clustering algorithm.

From: Construction of a predictive model for blood transfusion in patients undergoing total hip arthroplasty and identification of clinical heterogeneity

Clinical Characteristics

Overall

Cluster 1

Cluster 2

P-Value

(n = 61)

(n = 44)

(n = 17)

Age

   

 < 0.001

  Mean ± SD

68.84 ± 14.54

74.98 ± 8.81

52.94 ± 14.41

 

  Medium[P25,P75]

72 [61, 80]

74.5 [71, 82]

53 [46, 61]

 

BMI

   

 < 0.001

  Mean ± SD

21.82 ± 2.40

22.69 ± 2.11

19.56 ± 1.46

 

  Medium[P25,P75]

22.19 [19.83, 22.97]

22.61 [21.93, 23.39]

19.29 [18.44, 19.83]

 

HGB

   

0.250

  Mean ± SD

111.65 ± 17.26

110.28 ± 17.01

115.21 ± 17.40

 

  Medium[P25,P75]

111 [99, 122]

108.5 [98.75, 120.25]

118 [104, 129]

 

PLT

   

0.022

  Mean ± SD

263.23 ± 82.04

279.33 ± 83.84

221.56 ± 59.59

 

  Medium[P25,P75]

250 [212, 317]

273 [218, 335.5]

228 [188, 243]

 

Bleeding Volume

   

0.002

  Mean ± SD

509.84 ± 200.37

452.27 ± 146.53

658.82 ± 240.24

 

  Medium[P25,P75]

500 [400, 600]

475 [400, 525]

700 [500, 800]

 

Urine Volume

   

0.046

  Mean ± SD

764.43 ± 461.90

682.95 ± 374.31

975.29 ± 584.21

 

  Medium[P25,P75]

600 [450, 850]

575 [400, 800]

800 [600, 1300]

 
Back to article page