Estimators for Disease Dynamics with Imperfect Surveillance
Abstract
This report presents a method for estimating infective disease-dynamics parameters when contact rates are uneven, surveillance data are not systematically sampled, and cases are underreported. An important parameter for predictive infectious disease models is the effective reproductive number (Re). Re determines the rate at which new infections occur and respond to intervention strategies, such as vaccination, quarantine, and social distancing. However, accurate estimation of Re is complicated by shortcomings in surveillance data collection, and these shortcomings are difficult to mitigate through changes in sampling methods. The author proposes that estimation of Re is not necessary to model changes in disease dynamics; rather, the basic reproductive number R0 may be used along with contact parameters derived from network characteristics within the host population. In addition, estimates of R0 can be derived from imperfect surveillance data through application of hierarchical methods that correct for underreporting by using explicit estimates of detection probabilities. A hierarchical data-assimilative method for improving parameter estimates in predictive models when data are imperfectly collected is demonstrated in this report. Accurate estimates of changes in disease dynamics can inform management decisions and mitigation strategies.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 2021
- Accession Number
- AD1153457
Entities
People
- Tom Ingersoll
Organizations
- United States Army Soldier Systems Center