Comparison of methodologies for predicting
mobility performance in adults with low vision
PSW Fuhr, O.D., Ph.D., KH Chung, B.S., JL Elliott,
M.A., TK Kuyk, Ph.D.
Birmingham VA Medical Center, Birmingham, AL
Objectives: Better methods for serving increasing
numbers of visually impaired veterans may be devised with knowledge of
factors that contribute to performance of essential activities of daily
living. High in importance to functional independence is orientation to
and mobility within the environment. We have previously reported on visual
correlates of obstacle avoidance under photopic and mesopic conditions
in veterans with low vision (Kuyk, Elliott & Fuhr, 1998). The purpose
of this study was to determine if measuring and combining visual acuity
and visual field measures in the manner proposed in the 1999
Guide for the Evaluation of Visual Impairment (Guidebook) would better
predict mobility performance on an obstacle course than the measures of
visual function we used in our previous research.
Methods: Visual acuity, visual field, contrast
sensitivity, and scanning data from 156 legally blind veterans in our
previous mobility study (described in detail in previous publications)
were re-analyzed. Binocular visual fields were measured with Goldmann
kinetic perimetry, III4e stimulus at standard background luminance along
12 meridians. Amount of visual field remaining (visual field extend, VFE)
was determined by the sum of the 12 meridians, corrected for intervening
scotomas. The same visual field tests were re-scored by the methodology
proposed in the Guidebook to determine the Functional Field Score (FFS).
Monocular and binocular visual acuities that contribute to the Functional
Acuity Score (FAS) were obtained from the subject’s records, and these
were combined (multiplied) to determine the Functional Vision Score (FVS).
Mobility performance measures included time required to walk the course
and the total number of errors made during the walk. Correlation and regression
analyses were performed. Step-wise linear regression was used to compare
the strength of multi-predictor models composed of the new functional
scores (FFS, FAS and FVS) versus those composed of the measures of visual
function used in the original study.
Results: Of the two field (VFE & FFS)
and two acuity (high contrast and FAS) measures, only the field measures
were significantly correlated with performance. Combining FFS and FAS
to form FVS, did not improve correlations, but instead lowered them slightly
below those for FFS alone. In general the visual field measure from our
original study (VFE) was more highly correlated with the performance variables
that FFS. The best predictive models of time to complete and errors made
on the obstacle course under photopic and mesopic lighting conditions
combined VFE and scanning ability and these could account for upward of
approximately 40% of the variance in performance.
Conclusions: The method of combining field
and acuity measures as outlined in the Guidebook did not provide any advantage
in predicting mobility performance over either variable alone. If anything
it lowered correlations. This may be due to the fact FVS is obtained by
multiplying two variables, one of which (FAS) has almost no correlation
with performance. Future plans are to evaluate FVS as a predictor of other
outcome measures where both acuity and field may be important determiners
of performance.
Funding Acknowledgment: This study was funded
by the VA Rehabilitation Research and Development Service, project # C775-RA
and the VAMC, Birmingham, AL