Table 5.

Multivariable Impact and Contribution of Single Bilateral VFI on VRQoL

Exposure: Single VFI	β Coefficient* (95% CI)	P Value	% Change	Standardized Dominance Statistic†
Outcome: VRQoL overall
VAI
No	Reference	NA	NA
Yes	−0.34 (−0.54, −0.14)	<0.001	−8.43	17.6
CSI
No	Reference	NA	NA
Yes	−0.33 (−0.50, −0.17)	<0.001	−8.28	24.9
CVI
No	Reference	NA	NA
Yes	−0.20 (−0.43, 0.04)	0.100	−4.93	4.1
DPI
No	Reference	NA	NA
Yes	−0.25 (−0.40, −0.10)	0.001	−6.27	15.0
Outcome: Visual Functioning
VAI
No	Reference	NA	NA
Yes	−0.28 (−0.42, −0.13)	<0.001	−7.69	20.4
CSI
No	Reference	NA	NA
Yes	−0.23 (−0.35, −0.12)	<0.001	−6.55	24.0
CVI
No	Reference	NA	NA
Yes	−0.21 (−0.38, −0.04)	0.013	−5.94	7.6
DPI
No	Reference	NA	NA
Yes	−0.17 (−0.28, −0.07)	0.002	−4.94	15.1
Outcome: Emotional Well-being
VAI
No	Reference	NA	NA
Yes	−0.29 (−0.52, −0.07)	0.012	−5.83	13.5
CSI
No	Reference	NA	NA
Yes	−0.30 (−0.49, −0.12)	0.002	−6.03	20.9
CVI
No	Reference	NA	NA
Yes	−0.05 (−0.32, 0.21)	0.711	−1.03	0.8
DPI
No	Reference	NA	NA
Yes	−0.30 (−0.48, −0.13)	<0.001	−6.07	18.4

Minimal clinically important difference (defined as 0.5 SD of baseline VRQo) = 0.63. All models were adjusted for age, gender, ethnicity, socioeconomic status, living alone, smoking status, alcohol frequency, body mass index, diabetes, hypertension, dyslipidemia, CVD, CKD, and polypharmacy. Furthermore, each of the single VFI models was mutually adjusted for the other VFIs. Adjusted R² values of the models with overall, visual functioning and emotional well-being scores of VRQoL are 0.0593, 0.0703, and 0.0376, respectively.

^*Coefficients derived from linear regression models.

^†Standardized dominance statistics are expressed in percentages