Table 4.

Adjusted¹ prevalence ratios (PRs) and 95% CI for EQ-5D impairments for total sample

EQ-5D impairments	Model 1	Model 2	Model 3
	PRs (95% CI)	PRs (95% CI)	PRs (95% CI)
Mobility
Hui (ref. Han)	1.19 (0.76–1.87)	1.12 (0.71–1.77)	0.95 (0.61–1.47)
Low HL (ref. high HL)		1.74 (0.97–3.13)	1.67 (0.94–2.98)
Self-care
Hui (ref. Han)	1.18 (0.66–2.10)	1.2 (0.67–2.15)	1.01 (0.57–1.79)
Low HL (ref. high HL)		0.84 (0.42–1.69)	0.81 (0.41–1.59)
Usual activities
Hui (ref. Han)	0.90 (0.65–1.26)	0.90 (0.65–1.26)	0.80 (0.58–1.11)
Low HL (ref. high HL)		1.0 (0.66–1.51)	0.96 (0.64–1.43)
Pain/discomfort
Hui (ref. Han)	1.23(1.04–1.46)	1.21 (1.02–1.44)	1.02 (0.89–1.18)
Low HL (ref. high HL)		1.23 (1.01–1.50)	1.13 (0.95–1.33)
Anxiety/depression
Hui (ref. Han)	1.31 (1.06–1.63)	1.31 (1.05–1.62)	1.17 (0.95–1.44)
Low HL (ref. high HL)		1.07 (0.83–1.38)	1.04 (0.82–1.32)

1: Log-binomial regression models. Model 1 introduced socio-demographics (age, ethnicity, education, income, geographic location). Model 2 further introduced HL based on model 1. Model 3 further introduced the presence of chronic diseases based on model 2.