Table 2

Odds Ratio models by patient group, gender, age and education
Importance Experience
OR's model a OR's model b LR test OR's model a OR's model b LR test
Shared decision making p < 0.05 p < 0.05
…Benign breast abnormality reference reference reference reference
…Malignant breast …abnormality 1,03 1,02 1,28 1,49
…Reumatoid arthritis 0,87 0,87 1,54 1,75
…Spinal disc herniation 1,46 1,39 1,38 1,65
…Female reference reference reference reference
…Male 0,48 0,50 1,01 0,96
…Age < 45 reference reference reference reference
…Age 45-65 1,71 1,77 1,17 1,14
…Age > 65 1,45 1,59 1,18 1,07
…High education reference reference reference reference
…Medium education - 0,92 - 0,83
…Low education - 0,67 - 1,34
Understandable explanations1 n.s. n.s.
…Benign breast abnormality reference reference reference reference
…Malignant breast …abnormality 0,77 0,77 1,08 1,09
…Reumatoid arthritis 0,37 0,38 0,95 0,96
…Spinal disc herniation 0,56 0,56 1,11 1,11
…Female reference reference reference reference
…Male 0,65 0,66 1,66 1,64
…Age < 45 reference reference reference reference
…Age 45-65 1,95 1,97 1,43 1,42
…Age > 65 1,55 1,61 1,28 1,26
…High education reference reference reference reference
…Medium education - 0,95 - 0,90
…Low education - 0,84 - 1,00
Being able to ask questions2 p < 0.05 n.s.
…Benign breast abnormality reference reference reference reference
…Malignant breast …abnormality 0,91 0,93 1,86 1,85
…Reumatoid arthritis 0,57 0,60 2,06 2,03
…Spinal disc herniation 0,62 0,61 1,13 1,14
…Female reference reference reference reference
…Male 0,59 0,63 1,59 1,56
…Age < 45 reference reference reference reference
…Age 45-65 1,40 1,47 1,44 1,43
…Age > 65 0,92 1,07 1,27 1,22
…High education reference reference reference reference
…Medium education - 0,83 - 1,06
…Low education - 0,53 - 1,19

significant results (p < 0.05) are bold.

model a: patient group, gender and age.

model b: model a + education.

LR test: likelihood ratio test.

1 The proportional odds assumption was violated for the experience scores. A subsequent analysis that did not depend on the proportional odds assumption yielded virtually identical results regarding the transfer from 'usually' to 'always' and showed some deviations for the transfer from 'never / sometimes' to 'usually'.

2 The proportional odds assumption was violated for the importance scores. A subsequent analysis that did not depend on the proportional odds assumption yielded virtually identical results regarding the transfer from 'important' to 'of the utmost importance' and showed some deviations for the transfer from 'not (really) important' to 'important'.

Rademakers et al.

Rademakers et al. BMC Health Services Research 2012 12:261   doi:10.1186/1472-6963-12-261

Open Data