Aims To develop item banks for a patient-reported outcome measure (PROM) specific to vitreous floaters using input from patients. We report on the content generation and item refinement, and compare the content of the newly-developed PROM with published literature. Methods Potential PROM items (questions) were identified from two sources: 20 existing floaters-specific PROMs and two qualitative studies in patients with floaters. This initial item pool was evaluated with binning (grouping) and winnowing (reduction) to group the PROM items into quality-of-life domains. Patients with floaters provided feedback on the pilot PROM item banks in cognitive interviews. The Dutch PROM item banks were translated into English, Arabic and Turkish. Each step was guided by an expert panel consensus. Results The initial item pool with potential PROM items consisted of 921 items. After three rounds of binning and winnowing, the item pool was reduced to 272 items. After 10 cognitive interviews, 19 items were changed, 8 were deleted and 8 were added. The final PROM item pool consists of 272 items across 12 item banks (quality-of-life domains): visual symptoms, ocular symptoms, general symptoms, activity limitations, driving, mobility issues, health concerns, economic impact, emotional well-being, social well-being, inconvenience and coping. Conclusion The new floaters-specific PROM item banks included all 81 items from previous PROMs, supplemented with additional quality-of-life issues reported by people experiencing floaters (70.2%). The largest quality-of-life domains were ‘health concerns’ and ‘inconvenience’, emphasising the disease burden of experiencing vitreous floaters and patients’ information needs. Item banking allows clinicians and researchers to choose what domains and items to use in their measurement. Additional quality-of-life issues identified in previously unstudied populations can be added and calibrated with the existing items. Future studies using the item banks can compare the quality-of-life impact of different clinical subgroups, and control for important confounders.
Jong et al. (Thu,) studied this question.