Development and validation of a novel Barrett ' s oesophagus patient reported outcome measure (B-PROM)

Background Patients with Barrett's oesophagus (BO) carry significant cancer worry, burden of symptoms, and lack disease-specific knowledge. Currently there is no validated BO patient reported outcome measure (PROM) to measure these factors for use in clinical practice and research, hence the aim of this study was to devise a novel, validated BO-specific tool, B-PROM. Methods Literature review, quantitative and qualitative research informed the initial item generation. The item bank was refined through a modified Delphi process between May and August 2021. The PROM was then tested through cognitive interviews and validated via multicentre testing between September 2021 and February 2023 with the aim to create a succinct tool which addresses the key important factors to BO patients and has strong psychometric properties. Findings B-PROM covers key themes of disease-specific knowledge, trust in clinicians, burden of symptoms, cancer worry and burden of surveillance. Validation results from 387 participants (response rate 40.8%) showed 93.3% of participants completed >95% of B-PROM. All individual items scored a completion rate of >95%. Mean completion time was 5 mins 34s for a sample group. Nineteen items showed a ceiling effect, 3 items showed a floor effect. Internal consistency overall demonstrated a Cronbach Alpha of 0.846, while predetermined subsections showed Cronbach alphas of 0.335, 0.718, 0.736, and 0.896. Inter-item analysis found 2 pairs of items with strong correlation, with only 6 items correlating weakly. Item-total correlation showed 19 items correlated well. Exploratory Factor analysis (EFA) with principal component analysis produced 5 components with Eigenvalues >1 of which 4/5 had satisfactory Cronbach alphas. Test-retest reliability showed no significant differences across single and average measures (p <= 0.001). Interpretation B-PROM is the first BO-specific PROM to be systematically evaluated. Validation findings show strong internal consistency, short completion time, low missingness and excellent test-retest reliability.