Measurement properties of a patient-reported outcome scale for peptic ulcer: a validation study using classical test...

To evaluate the measurement properties of the Patient-Reported Outcome Scale for Peptic Ulcer (PROISCD-PU) and to examine its reliability, validity, responsiveness, and domain-level measurement precision using an integrated framework of Classical Test Theory (CTT) and Generalizability Theory (GT).

A total of 300 patients with peptic ulcer were recruited from tertiary hospitals and completed the PROISCD-PU. Classical Test Theory was used to evaluate internal consistency, construct validity, clinical discriminant validity, and responsiveness. Generalizability Theory was applied to identify sources of measurement variance and estimate relative generalizability (Eρ²) and absolute dependability (Φ) coefficients under the observed and alternative item configurations. Because the GT analyses were conducted separately for each domain, the findings were interpreted as evidence of domain-level measurement precision.

The PROISCD-PU demonstrated acceptable internal consistency for most domains and satisfactory construct validity, supported by item-domain correlation analyses and structural equation modeling. Clinical discriminant validity was partially supported, with significant differences after Holm-Bonferroni correction observed in the spirituality/belief health domain, general module subtotal, and disease-specific domain. Significant improvements were observed following treatment across all domains and the total scale, with standardized response mean values indicating moderate to large responsiveness. Generalizability analyses showed that person-by-item interaction was the major source of measurement variance. Absolute dependability (Φ) was acceptable for the mental health, social health, and disease-specific domains, whereas comparatively lower Φ coefficients were observed for the physical health and spirituality/belief health domains, indicating priorities for future refinement.

The PROISCD-PU demonstrated generally acceptable reliability, validity, responsiveness, and domain-level measurement precision for assessing patient-reported outcomes in peptic ulcer disease. The combined use of Classical Test Theory and Generalizability Theory provides complementary psychometric evidence supporting its application in clinical outcome evaluation and quality-of-life research. Future studies should evaluate longitudinal score stability in clinically stable patients and apply multivariate Generalizability Theory to optimize the reliability and scoring structure of the total composite score.