Improving Estimation of Intervention Impact on Antibody Titer Increase in the Presence of Missing Values: A Proposed Method for Addressing Limit of Detection Issues
Year: 2022
Type: preprint
Abstract: In many laboratory assays that measure immunological quantities, a portion of the measured values fall below a limit of detection (LOD). This is also the case for the hemagglutination inhibition assay (HAI), a common method used to quantify antibodies in influenza research. The conventional approach is to treat values below the LOD as either equal to the LOD or LOD/2, which can introduce potential biases. These biases can become more pronounced when calculating compound measures such as the difference between post-vaccination and pre-vaccination antibody titers (titer increase). To address this issue, we conducted simulations using LOD measurements with LOD/2 values as the standard imputation. We then developed a new method to adjust coefficient estimates that account for the censored nature of measurements below the LOD. Applying this new method to data from an influenza vaccine cohort study, we compared the impact of vaccine dose on the titer increase of HAI. Author Summary Analysis of measurements obtained from widely used antibody assays frequently overlooks the underlying data structure, leading to potential biases in the results. To address this issue, we have developed a method that effectively reduces these biases.
Source: medRxiv
Authors: Yang Ge, W. Zane Billings, Amanda L. Skarlupka, Wannan Cao, Kevin K. Dobbin, Ted M. Ross, Andreas Handel, Ye Shen
Institution: University of Southern Mississippi
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4293572651