830-P: The Development of a New Computer Adaptive Test of Illness Burden for Use in People with Type 2 Diabetes Mellitus (T2DM)

Abstract

In recent years, a growing body of research has established that social determinants of health influence the health outcomes of patients, especially those with T2DM. We developed a new patient-reported outcome (PRO) measure of an important social determinant of behavior that is specific to persons with T2DM: illness burden. Qualitative and quantitative methods informed the development of this new PRO. A literature review and semi-structured key informant interviews guided the establishment of an initial item pool. Expert and patient review, reading level assessment (to ensure no item had >5th grade reading-level language), and a translatability review were all used to refine the original item pool to 47 items. The refined item pool was then tested in 225 patients with T2DM. No items had sparse response option cells or problems with monotonicity; two items were deleted due to low item-rest correlations. Factor analyses (exploratory and confirmatory) supported the retention of 29 items. With those 29 remaining items, a constrained (common slope) graded response model (GRM) fit assessment indicated that two items had misfit and were excluded. No items displayed differential item functioning by age, sex, education, or socio-economic status. Thus, the final item bank is comprised of 27 items. Psychometric considerations (based on item information values) were used, in conjunction with expert content review, to select a 6-item static short form. IRT-based item parameters were employed to make this new measure available as a computer adaptive test. Internal consistency of the new PRO was excellent (α ≥ .90 for all administration formats). The new Illness Burden item bank captures patients’ perceptions of the impact that having T2DM has on their daily lives. It can be used in conjunction with the broader measurement system, the Re-Engineered Discharge for Diabetes Computer Adaptive Test, to evaluate important social determinants of behavior in persons with T2DM. Disclosure S. Mitchell: Other Relationship; Self; See Yourself Health LLC, Speaker’s Bureau; Self; Merck Sharp & Dohme Corp. M. A. Kallen: None. A. Bragg: None. I. Moldovan: None. J. M. Howard: None. B. W. Jack: None. J. A. Miner: None. C. M. Graves: None. N. Carlozzi: None. Funding National Institute of Diabetes and Digestive and Kidney Diseases (1R21DK121092)