Revisiting patient satisfaction following total knee arthroplasty: a longitudinal observational study

Table 4 Final multivariable model results

Variable	Time	Estimate	SE	p	OR¹ (95%CI)
Predictors of the intercept
SF-MCS	1	0.118	0.034	0.001	3.25 (1.67–6.34)*
SF-MCS	2–1	0.01	0.026	0.686	1.11 (0.66–1.84)
SF-PCS	1	0.115	0.038	0.002	3.16 (1.50–6.65)*
SF-PCS	2–1	0.026	0.036	0.473	1.30 (0.64–2.63)
WOMAC PAIN²	1	0.418	0.111	< 0.001	2.65 (1.76–4.01)*
WOMAC PAIN²	2–1	0.488	0.105	< 0.001	2.31 (1.49–3.56)*
Predictors of the slope
SF-MCS	3–2	0.026	0.011	0.014	1.30 (1.05–1.61)*
SF-PCS	3–2	0.044	0.016	0.005	1.55 (1.13–2.12)*
WOMAC PAIN²	3–2	0.109	0.034	0.001	1.24 (1.09–1.42)*
Correlation intercept and slope		−1.161	2.184	0.595
slope (mean)		−0.334	0.433	0.441
intercept (threshold)		2.244	2.384	0.346
Residual variance: intercept		8.851	9.003	0.326
Residual variance: slope		0.159	0.417	0.703

Notes Time 1 is baseline, Time 2 is 6 months, and Time 3 is 12 months. Log likelihood − 227.67, BIC 548.03, ¹OR odds ratio pertaining to a 10% difference in scores of the continuous independent variables comparing the odds of being in a higher satisfaction category relative to a lower satisfaction category. ²Reverse-scored, such that higher score indicates less pain

ISSN: 1471-2474