An item-level response shift study on the change of health state with the rating of asthma-specific quality of life: a report from the PROMIS(®) Pediatric Asthma Study.

PURPOSE: To examine item-level response shift associated with the change in asthma-related health state (i.e., change in asthma control status and global rating of change (GRC) in breathing problems). METHODS: Study sample comprised 238 asthmatic children who were between 8 and 17.9 years and completed the Pediatric Asthma Quality of Life Questionnaire (PAQLQ) symptoms, emotion function, and activity limitation domains at baseline and a follow-up assessment. Structural equation modeling was implemented to assess item-level response shift associated with the change in asthma-related health state with the adjustment for the influence of confounding variables. The magnitude of item-level response shift and its influence on the change of domain scores was estimated using Cohen's effect sizes. RESULTS: We found no instances of item-level response shift. However, two items were identified with measurement bias related to GRC due to breathing problems. Specifically, asthmatic children with better/about the same GRC due to breathing problems reported lower scores for one item in the emotional domain at follow-up compared to those with deteriorated GRC due to breathing problems. In addition, asthmatic children with better/about the same GRC due to breathing problems reported better scores for another item in the symptom domain at baseline compared to those with deteriorated GRC due to breathing problems. The impact of measurement bias was small and did not bias the change of domain scores over time. CONCLUSIONS: No item-level response shift, but two instances of measurement bias, appears in asthmatic children. However, the impact of these measurement issues is negligible.

Duke Scholars

Author Bryce B. Reeve Population Health Sciences