Online Program Home
My Program

Abstract Details

Activity Number: 467 - Modeling, Design Strategies and Assessment of Biomarkers
Type: Contributed
Date/Time: Wednesday, July 31, 2019 : 8:30 AM to 10:20 AM
Sponsor: Biopharmaceutical Section
Abstract #304186 Presentation
Title: A New Method for the Analysis of Categorical Data with Repeated Measurements - Demonstrated by Precision Data Analysis for Clinical Diagnostics
Author(s): Tinghui Yu*
Companies: AstraZeneca
Keywords: Inter-Rater Agreement; APA; ANA; GLMM; Random Effects; D-statistic

Precision (variation) is a very important statistical characteristic of a clinical assay. The FDA requires strong evidence demonstrating precision of a clinical assay to support its approval. To assess the precision of an assay, one needs to perform a study testing a set of representative samples under different conditions. This study design usually leads to repeated measurement data clustered by multiple levels of control factors. None of the existing statistical methods can provide satisfactory inference regarding the precision of a qualitative assay. We proposed a new method, based on a “D-statistic” to address this issue. The D-statistic is a normalized index measuring the random effect regarding each individual control factor. It defines a partition of the total variation in the test results. There is an intuitive probabilistic interpretation for our new method, which can facilitate the definition of unambiguous and quantifiable decision rules about the test performance. We also showed that the D-statistic is closely connected to GLMM. It can be easily generalized to offer a brand new prospective and additional inference power to the mixed effect models.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program