Evaluating Diagnostic Utility from the Literature

• Activity Number: 233
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): John Johnson*+
• Companies: PPD,
• Keywords: diagnostic utility ; metaanalysis ; ROC ; AUC ; binormal

Abstract

To assess the potential value of a new diagnostic device, a method for evaluating the diagnostic utility of a biomarker from literature reports is explored, with P300 (a fluctuation in brain voltage occurring approximately 300 ms after a stimulus) in post-traumatic stress disorder (PTSD) and non-PTSD subjects used as an example. Most relevant articles report, at most, a difference in the P300 between two groups. While meta-analytic methods of the literature may be used to summarize the distribution of this difference, it says little about diagnostic utility. We present a method of using the proper binormal model of receiver-operating characteristic (ROC) curves to use the meta-analysis to directly address the diagnostic utility of a medical device that measures P300. Though limitations of the method may prevent it from making a definitive assessment, such information may be used to evaluate the utility of developing a diagnostic product or design studies focused on diagnostic utility.