Developing More Precise and Meaningfully Discriminative Measures of Health-Related Quality of Life: A Case for Computer-Adaptive Testing

• Activity Number: 8
• Type: Invited
• Date/Time: Sunday, July 31, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Consulting
• Author(s): Laura Lee Johnson*+
• Companies: NCCAM
• Keywords: computer adaptive testing ; PROMIS ; psychometrics ; item banks ; patient reported outcome measurement ; scale development

Abstract

Many statisticians focus on the use of a scale as an outcome instrument an investigator has chosen for a research study. Better understanding how those scales are developed helps statisticians improve the selection of study outcomes to better fit a study's design, specific aims and population and also provides a basis for involvement of a statistician in new scale development. Many statisticians approached with measure or scale development questions have not had training in this area. Other speakers will describe classical and modern theory techniques including item response theory. This talk will focus on an example of a state-of the art application of several scale development methods including evaluating data quality, model assumptions and fit, and item calibration methods within the context of the new NIH funded Patient-Reported Outcomes Measurement Information System (PROMIS®). This project was challenged with creating psychometrically sound and valid item banks in a short amount of time and continues to build on its initial project. We will discuss briefly how PROMIS develops item banks, short forms, and how items are selected in computer adaptive testing (CAT).