Patient-reported outcomes (PROs) are increasingly accepted as important outcomes. We seek to extensively advance outcome assessment in arthritis and aging by developing improved instruments using item response theory (IRT) and computerized adaptive testing (CAT). The use of existing tools has been limited by impracticality (respondent burden) and difficulties comparing study results across tools. Shorter questionnaires have proven more practical, but """"""""ceiling"""""""" and """"""""floor"""""""" effects typically result from measuring a limited range within each health domain. Further, brief questionnaires (e.g., the SF- 36, the HAQ) lack the precision necessary to reliably detect changes in individual patient scores. To these issues, we bring a team with extensive experience in developing, evaluating, and implementing the most frequently used PRO assessment instruments in rheumatic diseases, large longitudinal patient datasets in musculoskeletal disease and aging, and experience with innovative approaches to the application of IRT and CAT. We propose to realize eight specific aims related to item bank development, psychometric testing, and CAT development, with a goal of delivering psychometrically sound tools to facilitate NIH clinical research. This project will build on the substantial infrastructure of Arthritis, Rheumatism, and Aging Medical Information System (ARAMIS) longitudinal data banks for item development, IRT testing, and CAT simulation and development. It will culminate in a four-ann randomized trial comparing enhanced traditional static and dynamic tools in terms of relative sensitivity to change. The first network project investigates the conceptual hierarchies of PROMIS domains and the utility of aggregations of domains seeking consensus, combining generic and disease-specific attributes, and envisioning a single generic/specific CAT application across diverse chronic conditions. The second network project examines the impact of survey mode on PROs and seeks to understand which interface options (e.g., telephone interview, telephone Interactive Voice Response, web-based desktop computer, and handheld computer) are most accessible across types of impairment and other stratified variables. Proposed iterative randomized trials seek to elucidate how administration mode influences dynamic PROs. Item banks for musculoskeletal diseases, aging, and other conditions, and improvement of responsiveness, reliability and validity will better represent the patient's perspective and enable more cost effective clinical research. Key Words: Rheumatic, Musculoskeletal, Aging, Health Outcomes, Computer Adaptive Testing (CAT), Item Response Theory (IRT).
Showing the most recent 10 out of 104 publications
