Back to search

Boston University

Bayesian methods for incorporating migration into small area estimation of infectious disease burden

Abstract

dc:description.abstract

Migration, or long-term relocation of residence, can be a driver in the transmission and spread of infectious diseases, particularly those with long latent periods (the time between infection and becoming infectious) or infectious periods (the time during which an individual can transmit the disease to others) (1,2). Reliable estimation and prediction of disease burden therefore require methods that explicitly account for migration. This is especially important for small areas – typically at least two administrative levels below the nation, such as counties or districts – where data are sparse, geographical disparities are more pronounced, and migration can quickly shift disease dynamics. Existing methods for incorporating migration in small area estimation either require infeasibly detailed data or rely on unrealistic simplifying assumptions. The focus of this dissertation is to develop methods to incorporate migration into the estimation and prediction of infectious disease metrics in small areas, while accommodating the asymmetric and directional nature of migration patterns and the often-limited availability of infectious disease data. This dissertation proposes two complementary methods for incorporating migration into small area estimation and prediction, with applications both to the binomial and multinomial settings. The first approach introduces a Bayesian power prior framework to adjust baseline estimates of infectious disease burden for migration. Power priors are most commonly used in clinical trial settings to incorporate historical data; here, they are repurposed to incorporate information from migrant origin areas to produce migration-adjusted estimates. The second approach incorporates a migration-informed matrix into the spatial random effect terms of existing Bayesian small area estimation models. Simulation studies demonstrate that both methods improve aspects of estimation relative to migration-naïve methods, with performance varying by migration intensity and structure. The results highlight the importance of explicitly accounting for migration in small area estimation and prediction of infectious disease burden. These methods offer distinct advantages and trade-offs, making them suitable to different settings and applications. Both approaches preserve the asymmetry of migration and do not substantially increase data requirements for generating small area estimates and predictions.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy
Level thesis:degree_level
doctoral
Discipline thesis:degree_discipline
Biostatistics
Grantor
Boston University
Year dc:date.issued
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Dupuis, Genevieve
Advisor dc:contributor.advisor
  • Jenkins, Helen
ORCID dc:identifier.orcid

Subjects

dc:subject × 6

Rights

dc:rights
Statement dc:rights
  • Attribution-NonCommercial-NoDerivatives 4.0 International
Language dc:language.iso
en_US

Identifiers

dc:identifier.*
Handle dc:identifier.uri
https://hdl.handle.net/2144/53596

Chain of custody

source
Harvested from
Boston University
Base URL
open.bu.edu/oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Dupuis, Genevieve. Bayesian methods for incorporating migration into small area estimation of infectious disease burden. doctoral thesis, Boston University, 2026. https://hdl.handle.net/2144/53596