On Statistical Properties of a Content Validity-Based Bayesian Approach for Validating Measures of Patient-Reported Outcomes

• Activity Number: 288
• Type: Topic Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Social Statistics Section
• Author(s): Byron J. Gajewski*+ and Valorie Coffland and Jamie Leopold and Nancy Dunton and Diane Boyle and Marge Bott
• Companies: The University of Kansas and The University of Kansas School of Nursing and The University of Kansas School of Nursing and The University of Kansas School of Nursing and The University of Kansas School of Nursing and The University of Kansas School of Nursing
• Keywords: Bayesian Analysis ; Equivalency ; NDQI ; Psychometrics

Abstract

As evolving healthcare therapies, treatments, and policies are introduced, it is more critical than ever to quickly develop new measures of patient reported outcomes (PRO). However, it is a lengthy process taking several years. To achieve more efficient and economical validity evidence, it has been proposed to integrate content validity expert and participants' data via a Bayesian approach. The purpose of this paper is to study statistical properties of this Bayesian approach using (a) asymptotic theory and (b) data collected from the National Database of Nursing Quality Indicators® (NDNQI®). The Bayesian approach is very promising. We conclude from this study that if future investigators use the Bayesian approach to develop measures of PRO they can reduce participant numbers by 37%, significantly decreasing level of effort and time.