Effects of a medical second opinion programme on patients’ decision for or against knee arthroplasty and their satisfaction with the programme

Table 4 Multivariable logistic regression model, predictors for recommendation of knee arthroplasty by the second opinion physician

Criterion	Unadjusted odds ratio	Multivariable logistic regression
Criterion	Unadjusted odds ratio	Adjusted odds ratio^d	95% CI		p-value
Kellgren-Lawrence score^a (3 or 4 versus 1 or 2)	16.09	17.24	3.65	81.38	0.001
KOOS^b score – pain	1.01	0.99	0.95	1.02	0.486
KOOS^b score – quality of life	1.03	1.04	1.00	1.08	0.041
Exercise therapy^c	1.89	1.88	0.63	5.57	0.258
Pain medication (reference: none)					0.089
On demand	0.73	0.38	0.12	1.21	0.101
Regular	1.50	1.10	0.31	3.91	0.873
Age – yr	1.02	1.03	0.98	1.08	0.234
Sex (reference: male)	1.09	1.02	0.41	2.53	0.975
Nagelkerkes R²: 0.346^e

^aKellgren-Lawrence score: range from 0 to 4, with higher scores indicating more severe disease. Scores of 1 and 2 as well as scores of 3 and 4 were combined
^bKOOS: Knee Injury and Osteoarthritis Outcome Score, scores ranging from 0 (worst) to 100 (best)
^cExercise therapy: supervised strength or endurance training (individually or in group), home-based strength or endurance training or comprehensive rehabilitation intervention
^dAdjusted odds ratio, meaning: Kellgren-Lawrence score^a: 17.24 times higher probability for a recommendation of grade 3 or 4 than of grade 1 or 2; KOOS Quality of Life: Probability of a recommendation increases by 4% per unit increase;
^eNagelkerkes R²: The explained variance of the model was 34.6%

ISSN: 1471-2474