Sensitivity and Uncertainty Analysis for Capture-Recapture Methods in Disease Surveillance

• Activity Number: 414 - Risk Modeling and Regression Techniques
• Type: Contributed
• Date/Time: Thursday, August 12, 2021 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Yuzi Zhang* and Jiandong Chen and Lin Ge and Lance Waller and Robert H Lyles
• Companies: Department of Biostatistics and Bioinformatics, RSPH of Emory University and Department of Biostatistics and Bioinformatics, RSPH, Emory University and Emory University and Emory University and Emory University
• Keywords: sensitivity analysis; uncertainty analysis; capture-recapture

Abstract

Capture-recapture (CRC) methods are widely applied in estimating the number (N) of incident cases in disease surveillance. We propose a sensitivity and uncertainty analysis based on the maximum likelihood estimation procedure that hinges on a key inestimable parameter for CRC experiments with an arbitrary number of surveillance efforts. Using the proposed sensitivity analysis focusing on the key inestimable parameter based on HIV surveillance data, we emphasize the importance of admitting the lack of information in the observed data about that crucial parameter. The proposed uncertainty analysis is a simulation-based approach designed to more realistically acknowledge variability in the estimated N associated with uncertainty in an expert's opinion about the inestimable parameter together with the statistical uncertainty. Additionally, we demonstrate that the proposed uncertainty analysis can facilitate a general interval estimation approach to accompany CRC analyses. Extensive simulation studies illustrate the reliable performance of the proposed uncertainty analysis in quantifying uncertainties in estimating N, and in serving as an alternative interval estimation procedure.