Adjusting Incidence Estimates with Laboratory Test Performances: A Pragmatic Maximum Likelihood Estimation-Based Approach
Yingjie Weng,Lu Tian,Derek Boothroyd,Justin Lee,Kenny Zhang,Di Lu,Christina P Lindan,Jenna Bollyky,Beatrice Huang,George W Rutherford,Yvonne Maldonado,Manisha Desai,Christina P. Lindan,George W. Rutherford,
DOI: https://doi.org/10.1097/ede.0000000000001725
2024-03-07
Epidemiology
Abstract:Understanding the incidence of disease is often crucial for public policy decision-making, as observed during the COVID-19 pandemic. Estimating incidence is challenging, however, when the definition of incidence relies on tests that imperfectly measure disease, as in the case when assays with variable performance are used to detect the SARS-CoV-2 virus. To our knowledge there are no pragmatic methods to address the bias introduced by performance of labs in testing for the virus. In the setting of a longitudinal study, we developed a maximum likelihood estimation (MLE)-based approach to estimate laboratory performance-adjusted incidence using the expectation-maximization algorithm. We constructed confidence intervals (CIs) using both bootstrapped-based and large-sample interval estimator approaches. We evaluated our methods through extensive simulation and applied them to a real-world study (TrackCOVID), where the primary goal was to determine the incidence of and risk factors for SARS-CoV-2 infection in the San Francisco Bay Area during July 2020 to March 2021. Our simulations demonstrated that our method converged rapidly with accurate estimates under a variety of scenarios. Bootstrapped-based confidence intervals (CIs) were comparable to the large-sample estimator CIs with a reasonable number of incident cases, shown via a simulation scenario based on the real TrackCOVID study. In more extreme simulated scenarios, the coverage of large-sample interval estimation outperformed the bootstrapped-based approach. Results from application to the TrackCOVID Study suggested that assuming perfect laboratory test performance can lead to inaccurate inference of the incidence. Our flexible, pragmatic method can be extended to a variety of disease and study settings.
public, environmental & occupational health