Relationship of Self-reported and Performance-based Visual Function With Performance-based Measures of Physical Function: The Health ABC Study

Abstract

Background: To assess the relationship between self-reported and performance-based visual impairment (VI) and lower extremity physical function.

Methods: Cross-sectional analysis of 2 219 Health ABC participants who completed vision testing and the Short Physical Performance Battery (SPPB). Linear regression models used either self-reported (weighted visual function question [VFQ] score) or performance-based (visual acuity [VA], log contrast sensitivity [LCS], Frisby stereoacuity [SA]) to predict SPPB or its components-gait speed, chair stands, or standing balance-with and without covariate adjustment.

Results: Mean age was 73.5 years (range 69-80); 52.4% were female and 37.4% African American. All VI measures were strongly associated with SPPB in unadjusted and adjusted models (p < .001). A self-reported VFQ score 1 standard deviation lower than the mean (mean 87.8 out of 100) demonstrated a -0.241 (95% confidence interval [CI]: -0.325, -0.156) adjusted difference in SPPB. After controlling for covariates, VA of <20/40 (41%) demonstrated a -0.496 (-0.660, -0.331) lower SPPB score while SA score>85 arcsec (30%) had a -0.449 (-0.627, -0.271) adjusted SPPB score versus those with better visual function. LCS < 1.55 (28.6%) was associated with a -0.759 (-0.938, -0.579) lower and LCS ≤ 1.30 (8%) with a -1.216 (-1.515, -0.918) lower adjusted SPPB score relative to better LCS. In a final multivariable model containing multiple vision measures, LCS remained independently associated with SPPB and all components, while SA remained associated with balance (all p < .05).

Conclusions: Both self-reported and performance-based VI are strongly associated with poor lower extremity physical function. These findings may identify a subgroup of older adults with co-existing visual and physical dysfunction who may benefit from targeted screening and intervention to prevent disability.

Keywords: Balance; Contrast sensitivity; Gait speed; Stereoacuity.

Figure 1.

Unadjusted bivariate and adjusted (all covariates except age) bivariate and multivariable models of vision variables and physical function: Differences (95% CI) in performance-based physical function outcomes comparing levels of discrete visual function predictors or for 1 standard deviation difference in continuous visual function predictors. Bivariate Unadjusted presents the unadjusted bivariate analysis for the association between each measure of visual function and physical function. Bivariate Adjusted presents the analysis for the association between each measure of visual function and physical function, adjusted for clinical and demographic covariates (sex, Black race, education, hypertension, diabetes mellitus, cardiovascular disease, cerebrovascular disease, BMI, CES-D score, smoking status, alcohol consumption). Multivariable Adjusted presents the multivariable analysis for the association of multiple measures of visual function (Frisby stereoacuity, log contrast sensitivity, logMAR visual acuity, and the weighted VFQ score) and each measure of physical function, adjusted for clinical and demographic covariates (sex, Black race, education, hypertension, diabetes mellitus, cardiovascular disease, cerebrovascular disease, BMI, CES-D score, smoking status, alcohol consumption.). The physical function outcomes are (A) SPPB, (B) gait speed (m/s), (C) chair stand pace (stands/s), and (D) balance (seconds). LCS = log contrast sensitivity; LogMAR = log₁₀ minimum angle of resolution of resolution; MV = multivariable; SPPB = Short Physical Performance Battery; VFQ = visual function questions.