Purchase this article with an account.
Robert W Massof, Judith E Goldstein, LOVRNET; Health State Effects on Accuracy and Precision of Visual Ability Measures in Low Vision Patients. Invest. Ophthalmol. Vis. Sci. 2014;55(13):4157.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The Activity Inventory (AI) is an adaptive rating scale questionnaire for measuring visual ability and its functional domains. Visual acuity is the strongest predictor of visual ability, but other health states also contribute to the measures. This study determines how physical functioning, depression, and cognitive impairment affect the accuracy and precision of low vision patient outcome measures made with the AI.
The AI, the Geriatric Depression Scale (GDS), the physical functioning component of the SF-36, and the Telephone Interview for Cognitive Status (TICS) were administered by telephone to 760 new low vision patients at 28 collaborating low vision centers throughout the U.S. prior to the patients' first visit appointment. The study was approved by the Johns Hopkins IRB and oral consent was obtained prior to the interview.
Rasch analysis of the AI item responses was used to estimate AI person measures, standard errors of the measure, and information weighted mean square fit statistics for each of the patients from their difficulty ratings of all valued items and of subsets of items that define four visual ability domains: reading, mobility, visual information processing, and visual motor function. Rasch analyses of GDS, SF-36, and TICS responses were employed to obtain measures of depression, physical functioning, and cognitive functioning from the same patients. Multivariate regression with health states as independent variables and AI person measures, standard errors of the measure, and information weighted mean squares as dependent variables showed that: 1) depressed mood biased visual ability measures and increased the standard error of the measure for all domains, but had no effect on mean squares; 2) physical functioning biased only mobility function measures and increased mean squares only for mobility, but had no effect on standard errors; 3) cognitive function biased reading function measures only and increased mean squares for reading and other visual ability domains, but had no effect on standard errors.
We conclude that 1) depression introduces response bias and decreases measurement precision for all visual ability domains ; 2) physical limitations affect the accuracy of mobility function measures, but have no effect on measurement precision; 3) cognitive disorders affect reading function accuracy, but, have no effect on measurement precision.
This PDF is available to Subscribers Only