Skip to main content
. 2018 Feb 21;8:3380. doi: 10.1038/s41598-018-21713-2

Table 3.

Logistic regression models identifying predictors of non-satisfaction 12-month after primary total knee replacement.

Intercept and Predictors (reference category) COASt (sample = 450) GAR (sample = 791)
Coefficient
(95% CI)
P value Re-estimated coefficient
(95% CI)
P value
Intercept −3.29 −0.68
Sex (Women)
  Men −0.19
(−0.82 to 0.44)
0.55 −0.22
(−0.63 to 0.18)
0.28
Age at operation, years 0.02
(−0.01 to 0.06)
0.23 −0.03
(−0.05 to −0.01)
0.003
Prior treatment for anxiety (No)
  Yes 0.997
(0.21 to 1.78)
0.01 0.05
(−0.53 to 0.63)
0.86
Current smoker (No)
  Yes 0.93
(−0.26 to 2.12)
0.12 0.41
(−0.12 to 0.94)
0.13
Injection of corticosteroids (No)
  Yes −1.04
(−1.87 to −0.21)
0.01 −0.22
(−0.78 to 0.34)
0.45
Standardised OKS/WOMAC, points −0.37
(−0.67 to −0.07)
0.02 −0.31
(−0.53 to −0.09)
0.006
AUC 0.69 0.64
Optimism 0.04
Bias-corrected AUC 0.65

Clinical outcomes in arthroplasty study, COASt; Geneva arthroplasty registry, GAR; 95% confidence intervals, CI; Oxford knee score, OKS; Western Ontario & McMaster Universities Osteoarthritis Index, WOMAC; Area Under the receiver operating characteristic Curve, AUC.