Self-reported assessment of visual function in a population-based study: the SEE project. Salisbury Eye Evaluation.

Abstract

PURPOSE: To report on the usefulness of the Activities of Daily Vision Scale (ADVS) questionnaire for assessing visual functioning, a population-based sample of elderly Americans. METHODS: The ADVS questionnaire was administered to a population-based sample of 2520 community-dwelling individuals 65 to 84 years of age in Salisbury, MD. Items and subscales were evaluated for internal consistency, item discrimination, and content validity. Published subscale groupings and item associations in our population were compared for coherence using correlation, factor, and cluster analyses. Whole-sample and race- and gender-specific analyses were conducted. External validity was explored by regressing ADVS scores on standard psychophysical vision measures. RESULTS: ADVS scores were skewed to high visual functioning levels; approximately 60% of the population had function scores of 95 or better (of a possible 100). The overall, night driving, and near vision scales were internally consistent and had strong item-subscale associations; the day driving and glare subscales were not acceptable regarding these properties. The far vision subscale was acceptably scalable but only weakly differentiated from the other subscales. Overall, night driving, near vision, and far vision scores were all statistically and independently associated with multiple psychophysical vision measures. Findings were consistent across race and gender subgroups. CONCLUSIONS: As assessed by the ADVS, reported visual functioning is high in our representative older population. The overall scale and selected subscales effectively distinguish persons along a spectrum of ability. They correlate with measures of visual impairment in a reasonable way and thus hold promise for risk factor investigations. The published day driving and glare subscales should be examined for relevance and consistency before being applied in population-based settings. Methods specific to population-based settings should be investigated for their ability to better elicit additional visual function dimensions and early visual disability.