Skip to main content
. 2018 Feb 21;8:3381. doi: 10.1038/s41598-018-21714-1

Table 2.

General linear model identifying predictors of knee pain and function 12-month after total knee replacement in the development dataset. OKS, Oxford knee score; IMD, index of multiple deprivation; BMI, body mass index; ASA grade, physical status scoring according to the American society of anaesthesiologists; ACL, anterior cruciate ligament; EQ-5D-3L, 3-level version of the EuroQol five dimensions questionnaire; CI, confidence interval. Variables included in the final regression model are those that are retained in at least 70% of the 200 bootstrap backward selection regression models.

Intercept and Predictors (reference category) Overall (n = 1,649)
% Retained in final model Coefficient (95% CI) P-value
Intercept 32.9
Age at operation (<60 years)
    60–69 years 100% (forced) 0.8 (−1.5 to 3.1) 0.477
    70 to 79 years 100% (forced) 1.4 (−0.9 to 3.6) 0.238
    80 years or more 100% (forced) −2.5 (−5.5 to 0.4) 0.088
Gender (Female)
    Male 100% (forced) −4.8 (−8.0 to −1.6) 0.003
Age x gender (<60 years x female)
    60–69 years x male 90% 4.8 (1.4 to 8.3) 0.006
    70 to 79 years x male 84% 4.3 (0.8 to 7.7) 0.015
    80 years or more x male 100% 8.1 (4.0 to 12.3) <0.001
IMD 2004 score (10 units) 97% −0.6 (−1.0 to −0.2) 0.008
BMI, kg/m2 (10 units) 96% −1.5 (−2.4 to −0.6) 0.001
EQ-5D-3L 5th question (No)
    Anxiety/depression 98% −1.6 (−2.5 to −0.6) 0.001
OKS baseline 100% 0.4 (0.3 to 0.4) <0.001
ASA grade (Fit and healthy)
    Asymptomatic no restriction 55%
    Symptomatic minimal/severe restriction 97% −2.0 (−3.2 to −0.8) 0.001
Previous knee arthroscopy (No)
    Yes 82% −1.6 (−3.0 to −0.2) 0.025
Other condition affecting mobility (No)
   Yes 100% −3.3 (−4.7 to −2.0) <0.001
Fixed flexion deformity (No)
    Yes 100% 1.7 (0.8 to 2.6) <0.001
Pre-operative ACL (Intact)
    Damaged/absent 81% 1.0 (0.1 to 2.0) 0.029
R 2 19.3%
Optimism 1.6
Bias-corrected R 2 17.6%