Abstract
Abstract: :
Purpose: To identify the content for a parsimonious, multi–attribute, vision–specific utility measure (VisQoL) to be used in conjunction with the generic quality of life instrument Assessment of Quality of Life (AQOL–II) to evaluate outcomes of eye care and rehabilitation programs. Methods: To develop the VisQoL scale, a database of 33 items was generated based on focus groups and previous research. Six domains were proposed: Physical well–being, Social well–being, Independence, Self–actualization, Emotional well–being, and Planning & Organization. For the pilot study, adult participants with a vision impairment (VA in better eye <20/30) were recruited from the Royal Victorian Eye and Ear Hospital (RVEEH), and adult participants with VA 20/20 or better in both eyes were recruited from a Centre for Eye Research Australia (CERA) database of previous research participants representing the general community. All participants completed VisQoL (33 items) and AQOL–II (20 items). To reduce the VisQoL item pool to the final structure, basic psychometric properties were examined. In addition, exploratory factor analyses (EFA), item response theory (IRT) analysis, and structural equation modeling (SEM) were conducted. Results: The item bank was administered to 374 participants; 39% were vision impaired. The mean age was 64.5 ± 11.6 years (range 22 – 94), with 53% female. From the 147 RVEEH participants, 31% had very mild vision impairment (<20/30 ≥20/40), 24% mild impairment (<20/40 ≥20/60), 32% moderate impairment (<20/60 ≥20/200), and 12% severe impairment (<20/200). Items were firstly eliminated based on basic psychometric properties, including items failing to discriminate between the two groups (normal vs impaired vision). Next, items were selected based on their EFA loading, with the 12 items from the final factor analysis loading on one factor at ≥0.68, with Cronbach α = 0.96. However, to minimize redundancy, IRT analysis and SEM were used to further reduce the VisQoL item pool to 6 items. The SEM statistics for the model indicate excellent fit properties (Χ2 = 6.359, df = 8; comparative fit index = 1.000, root mean square error of approximation = 0.000). Conclusions: Six items have been identified as suitable for the further development of a multi–attribute vision–specific measure that can generate utility values for the evaluation of vision–related programs and services.
Keywords: quality of life • clinical (human) or epidemiologic studies: health care delivery/economics/manpower • low vision