Introducing BREAST-Q Computerized Adaptive Testing
- Additional Document Info
- View All
BACKGROUND: The BREAST-Q is a widely used patient-reported outcome instrument measuring health-related quality-of-life and patient satisfaction in breast surgery. Shorter assessment potentially increases patients' willingness to complete scales, but simply offering a shortened version leads to unacceptable loss in measurement precision. The authors aimed to develop a computerized adaptive test (CAT) to shorten the BREAST-Q's Satisfaction with Breasts scale while maintaining reliability of measurement. METHODS: The authors created a CAT, which repetitively administered questions from the pool of 16 questions, until prespecified levels of reliability were reached [i.e., standard errors (SE) of 0.32 to 0.55]. In a simulation study, the authors tested the CAT's feasibility for all potential satisfaction scores. In a second study using actual patient data, 5000 breast reconstruction patients who had previously completed the full scale were randomly selected from a large database. Their full-scale satisfaction scores were compared with their CAT-derived scores. RESULTS: In both studies, by applying CAT, the Satisfaction with Breasts scale could be reduced to an average of 10 questions when using the minimum level of measurement precision for individual-patient measurement (SE, 0.32), compared with four questions when using the minimum precision level for group-based research (SE, 0.55). Score estimates were highly correlated between the CAT assessment and the full scale (0.91 to 0.98 in the simulation study, and 0.89 to 0.98 in the patient data study). CONCLUSIONS: Applying computerized adaptive testing to the BREAST-Q's Satisfaction with Breasts scale facilitates reliable assessment, with 38 to 75 percent fewer questions than the full version. The novel BREAST-Q CAT version may decrease response burden and help overcome barriers to implementation in routine care.
has subject area