Transportation of Diagnostic Accuracy Measures to a New Target Population

Principal Investigator

Name
Bing Li

Institution
Brown University

Position Title
student

Email
bing_li1@brown.edu

About this CDAS Project

Study
NLST (Learn more about this study)

Project ID
NLST-712

Initial CDAS Request Approval
Sep 20, 2020

Title
Transportation of Diagnostic Accuracy Measures to a New Target Population

Summary
When we have estimators from a population where the diagnostic test is evaluated (i.e. source population) but wish to evaluate the test on a different population of interest where the outcome of disease status is missing (i.e. target population), how to interpret and use the source population estimators in the context of target population (i.e. the problem of transportation) needs to be considered. Diagnostic accuracy can be varied with patients' demographics, geographic location, disease severity and etc. If the source population are not representative of the target population, bias occur, i.e. the estimates of test accuracy are not representative of the accuracy in the target population. Our goal is how to transport diagnostic accuracy measures from source population to target population.

Aims

1. Develop transportation methods for the area under ROC curve (AUC).
2. Extend the methods to the situation where the marker is missing for target population.
3. Apply the methods to transport AUC from NLST population to National Health and Nutrition Examination Survey (NHANES) population.

Collaborators

Jon Steingrimsson, Brown University
Constantine Gatsonis, Brown University

Related Publications

• Estimating the area under the ROC curve when transporting a prediction model to a target population.
  Li B, Gatsonis C, Dahabreh IJ, Steingrimsson JA
  Biometrics. 2022 Nov 17 PUBMED
• Transporting a prediction model for use in a new target population.
  Steingrimsson JA, Gatsonis C, Li B, Dahabreh IJ
  Am J Epidemiol. 2022 Jul 22 PUBMED