Purpose: The purpose of this study was to assess the psychometric properties of diabetic retinopathy (DR) and diabetic macular edema (DME) quality-of-life (QoL) item banks and determine the utility of the final calibrated item banks by simulating a computerized adaptive testing (CAT) application.
Methods: In this clinical, cross-sectional study, 514 participants with DR/DME (mean age ± SD, 60.4 ± 12.6 years; 64% male) answered 314 items grouped under nine QoL item pools: Visual Symptoms (SY); Ocular Comfort Symptoms (OS); Activity Limitation (AL); Mobility (MB); Emotional (EM); Health Concerns (HC); Social (SC); Convenience (CV); and Economic (EC). The psychometric properties of the item pools were assessed using Rasch analysis, and CAT simulations determined the average number of items administered at high and moderate precision levels.
Results: The SY, MB, EM, and HC item pools required minor amendments, mainly involving removal of six poorly worded, highly misfitting items. AL and CV required substantial modification to resolve multidimensionality, which resulted in two new item banks: Driving (DV) and Lighting (LT). Due to unresolvable psychometric issues, the OS, SC, and EC item pools were not pursued further. This iterative process resulted in eight operational item banks that underwent CAT simulations. Correlations between CAT and the full item banks were high (range, 0.88-0.99). On average, only 3.6 and 7.2 items were required to gain measurement at moderate and high precision, respectively.
Conclusions: Our eight psychometrically robust and efficient DR/DME item banks will enable researchers and clinicians to accurately assess the impact and effectiveness of treatment therapies for DR/DME in all areas of QoL.