Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 211 - Disease Prediction
Type: Contributed
Date/Time: Tuesday, August 10, 2021 : 1:30 PM to 3:20 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #318585
Title: Two-Stream Capture-Recapture Method with Accounting for Misclassification
Author(s): Lin Ge* and Robert H Lyles and Lance Waller and Yuzi Zhang and Timothy Lash and Kevin Ward
Companies: Emory University and Emory University and Emory University and Department of Biostatistics and Bioinformatics, RSPH of Emory University and Emory University and Emory University
Keywords: Capture-Recapture Method; Misclassification; Sensitivity Analysis; Uncertainty Analysis; Anchor Stream

Surveillance research is of great importance for monitoring disease prevalence among populations by effective statistical sampling approaches. Benefiting from multiple surveillance efforts in current epidemiological initiatives for monitoring disease, we extend and customize traditional capture-recapture (CRC) methods in tandem with epidemiological design and implementation principles in order to leverage error-prone diagnostic signals toward defensible estimates of case counts.

In this paper, we start with the independence assumption for two data streams (the LP, for "Lincoln-Petersen" condition) and propose a CRC framework to estimate case counts incorporating positive predictive values (PPV). Next, we relax the LP condition and assume the two data streams are not necessarily operating independently. In this case, we propose an approach for sensitivity and uncertainty analyses focusing on key non-identifiable parameters to generalize the LP condition-based CRC estimator. Finally, to justify the LP conditions by design, we consider the benefits of implementing an additional well-controlled surveillance effort (an “anchor stream”), which is conducted independent

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

Back to the full JSM 2021 program