Presented in Fort Lauderdale, Florida May 2005, sponsored by the Association for Research in Vision and Ophthalmology (ARVO). Published in Invest Ophthalmol Vis Sci 44: 3844.

K. Pesudovs¹, D.B. Elliott² and D.J. Coster¹

¹ Ophthalmology, Flinders University & Flinders Medical Centre, Adelaide, Australia
² Optometry, Bradford University, Bradford, United Kingdom

Commercial Relationships: K. Pesudovs, None; D.B. Elliott, None; D.J. Coster, None.

Support: NHMRC NHF Fellowship 0061

Abstract

Purpose: Conventionally developed and validated visual disability questionnaires do not result in true linear measurement of disability (Pesudovs Invest Ophthalmol Vis Sci 2003). The purpose was to rescale a conventional questionnaire, the Visual Disability Assessment (VDA, Pesudovs Br J Ophthalmol 1998), using Rasch analysis to create a linear measure of visual disability re–engineered to optimize item fit to the construct and to maximize targeting of item difficulty to cataract surgery patient ability. The psychometric performance of the new questionnaire was fully evaluated.

Methods: The 18–item VDA was administered to 270 consecutive patients presenting pre– or post– cataract surgery. This included 99 bilateral cataract, 115 second eye cataract and 56 pseudophakic patients. 124 patients had comorbid eye disease. The potential for improving the questionnaire through scale and item reduction was explored with Rasch analysis (Winsteps v3.49). This resulted in a new questionnaire: the Cataract Outcomes Questionnaire. Validity and reliability studies were performed. Responsiveness was determined on a further 122 patients undergoing cataract surgery. The minimal clinically important difference (MCID) was determined relative to a self–reported improvement in vision.

Results: Several VDA items poorly contributed to the measurement of visual disability: 8 items were removed, for 3 of the remaining 10 items the scale was reduced from 4 to 3 categories, and all were rescaled by Rasch analysis. The 10–item Cataract Outcomes Questionnaire had good precision, reliability and internal consistency (person separation, 2.05; reliability, 0.81; RMS measurement error, 6.27; mean square ± SD infit, 0.99 ± 0.45; outfit, 0.98 ± 0.54, item infit range 0.77 – 1.19 and item outfit range 0.66 – 1.31). The items (mean score 51.47 ± 8.54) were reasonably well targeted to the subjects (44.11 ± 5.91). Test retest reliability (intra–class correlation coefficient, 0.93; coefficient of repeatability, ± 3.70 units) and criterion validity (correlations with high (r=0.44) and low contrast visual acuity (r=0.44)) were sound. The responsiveness statistic for first and second eye cataract surgery was 7.75 and 4.60, and the MCID was 2.96 units (0–100 point scale).

Conclusions: The Cataract Outcomes Questionnaire quantifies the visual disability of cataract surgery patients and has been shown to be valid, reliable and responsive. Rasch scaling coefficients which convert categorical data into a continuous variable are supplied to simplify clinical application.

Key Words: treatment outcomes of cataract surgery • clinical (human) or epidemiologic studies: outcomes/complications • quality of life