Identifying the primary outcome for a randomised controlled trial in rheumatoid arthritis: the role of a discrete choice experiment

Table 5 Hierarchical importance of attributes based on partial log-likelihood analysis

Conditional logit model				Mixed logit model
Attribute level excluded from the analysis (by order of impact)	Log likelihood	Partial effect- change in log likelihood	Relative effect - %sum of change in log likelihood	Attribute level excluded from the analysis (by order of impact)	Log likelihood	Partial effect-change in log likelihood	Relative effect -%sum of change in log-likelihood
None	−1193.335			None	−1133.406
Pain_2	−1213.810	−20.475	0.276	Pain_2	−1190.501	−57.095	0.264
Pain_1	−1209.571	−16.236	0.219	Mobility_2	−1167.962	−34.556	0.160
Mobility_1	−1206.337	−13.002	0.175	DActivities_2	−1157.175	−23.769	0.110
DActivities_2	−1201.174	−7.839	0.106	Pain_1	−1156.766	−23.360	0.108
DActivities_1	−1197.358	−4.023	0.054	Mobility_1	−1155.712	−22.306	0.103
Fatigue_1	−1196.652	−3.317	0.045	DActivities_1	−1146.207	−12.801	0.059
Mobility_2	−1196.297	−2.962	0.040	Fatigue_1	−1144.621	−11.215	0.052
Shoes_1	−1196.081	−2.746	0.037	Shoes_1	−1144.179	−10.773	0.050
HF_3	−1194.546	−1.211	0.016	Shoes_2	−1142.456	−9.050	0.042
HF_2	−1194.473	−1.138	0.015	HF_1	−1138.126	−4.720	0.022
Shoes_2	−1194.413	−1.078	0.015	Fatigue_2	−1137.004	−3.598	0.017
Fatigue_2	−1193.553	−0.218	0.003	HF_2	−1135.117	−1.711	0.008
HF_1	−1193.386	−0.051	0.001	HF_3	−1134.987	−1.581	0.007

Notes:
_1 refer to “moderate” states
_2 refer to “perfect” or “no problems” states
Coefficients calculated based on the “extremely bad” state being the baseline
HF_1: ‘To improve my walking, I will go for supervised exercise as part of a group twice a week for 6 weeks’
HF_2: ‘To improve my walking, I will go for one-to-one supervised exercise, once a week for 12 weeks’
HF_3: ‘To improve my walking, I will go for one-to-one supervised exercise, twice a week for 6 weeks’;
Baseline is: ‘To improve my walking, I will go for supervised exercise as part of a group once a week for 12 weeks’

ISSN: 1757-1146