Constructing Item Banks for Measuring Quality of Life in Refractive Error

SIGNIFICANCE: Existing patient-reported outcome instruments in refractive error are paper-based questionnaires. They are not comprehensive and psychometrically robust. This study has identified the content of the refractive error-specific item banks that aim to provide comprehensive and scientific measurement of refractive error-specific quality of life. PURPOSE: The purpose of this study was to identify minimally representative, optimally informative, and efficient sets of items for measuring quality of life in people with refractive error. METHODS: First, items were identified from existing patient-reported outcome instruments. Second, items were developed from qualitative studies with people with refractive error (48 and 101 in-depth interviews in Australia and Nepal, respectively). Third, classification and selection of items were done based on a set of systematic criteria using an iterative process of binning and winnowing. The resulting items underwent cognitive testing with people with refractive error in Australia and in Nepal. Each step was guided by an expert panel consensus. RESULTS: We retrieved 792 items from the existing patient-reported outcome instruments. From the interviews conducted in Australia, a total of 2367 comments were coded into 807 initial items. Similarly, from the interviews conducted in Nepal, 3477 comments were coded into 914 initial items. After binning and winnowing, followed by cognitive testing, a final set of items comprising 337 items for the Item-pool (Australia) and 308 items for the Item-pool (Nepal), both spanning 12 domains, was obtained. Forty-seven percent of items were common across the two item pools. In the Item-pool (Nepal), 65% items were common for corrected and uncorrected refractive error. CONCLUSIONS: We identified the content of two different sets of item banks to comprehensively measure the impact of refractive error on quality of life for people in Australia and Nepal, which may be applicable to high-income country settings and low-and middle-incomecountry settings, respectively. Future work aims to develop computeradaptive testing system to administer the item banks, resulting in useful measurement tools for researchers, clinicians, and policy planners.