How do EQ-5D-3L and EQ-5D-5L compare in a Swedish total hip replacement population?

Figures & data

Figure 1. Preoperative and 1-year postoperative procedures for collecting the EQ-5D-3L and -5L questionnaires. ^aNon-respondents were reminded after 1 month according to regular procedure.

Figure 2. Possible redistribution of responses between EQ-5D-3L and -5L. Inconsistency = choosing an answer 2 levels from the first.

Table 1. Demographics pre- and 1 year postoperatively THR

Factor	n = 1,554
Age, mean (SD)	70 (11)
Female sex, n (%)	878 (57)
ASA, n (%)
1	340 (22)
2	967 (63)
3	229 (15)
4	1 (0.1)
BMI, mean (SD)	28 (4.8)
EQ VAS preoperatively, mean (SD)	57 (22)
EQ VAS postoperatively, mean (SD)	74 (21)
EQ-5D index preoperatively, mean (SD)	0.40 (0.3)
EQ-5D index postoperatively, mean (SD)	0.76 (0.3)

Table 3. Percentage of patients reporting severe problems (floor effect) and no problems (ceiling effect)

	Preoperative patients (n = 524)				Postoperative patients (n = 508)
	Severe problems (floor effect)		No problems (ceiling effects)		Severe problems (floor effect)		No problems (ceiling effects)
EQ-5D dimensions	3L	5L	3L	5L	3L	5L	3L	5L
Mobility	0	2	6	2	0	0.2	51	44
Self-care	0.6	0.6	67	30	0.2	0.2	91	76
Usual activities	11	9	27	4	2	1	71	46
Pain/discomfort	44	4	0.6	0.6	5	0.2	39	35
Anxiety/depression	3	0.4	53	39	2	0.2	73	69
Overall	0	0	1	0.4	0	0	32	25

Figure 3. Redistributions of answers between EQ-5D-3L and -5L for all patients.

Figure 4. Strength of association and agreement between the EQ VAS scores obtained with the EQ-5D-3L and -5L questionnaires approximately 2 weeks apart. The diameter of the points is proportional to the number of patients reporting that particular score. Note the different scaling of the y-axis in the Bland–Altman plots to improve visibility.