Back to search

University of Thessaly (UTH)

Application of Bayesian methods and advanced algorithms in the measurement of zoonotic disease occurrence

Abstract

dc:description

This doctoral thesis represents a holistic and comprehensive exploration of diagnostic test evaluation, prevalence estimation, substantiating freedom from disease, and disease surveillance & early outbreak prediction through advanced Bayesian statistical methodologies, focusing on zoonotic diseases. Zoonotic diseases pose significant challenges due to their complex interplay between humans, animals, and the environment. This work seeks to address these challenges by building a methodological “toolbox” that integrates advanced approaches for “measuring” disease occurrence, offering practical solutions for real-world public health and veterinary scenarios. Through the detailed examination of Q fever, brucellosis, and influenza as case studies, the thesis highlights how advanced statistical tools can be applied to tackle critical issues in zoonotic disease management. The research begins with the foundational step of diagnostic evaluation, focusing on the accuracy and reliability of tests. By employing Bayesian Latent Class Models (BLCMs), the thesis demonstrates how to estimate the sensitivity (Se) and specificity (Sp) of diagnostic tools in the absence of a gold standard, addressing key challenges such as sparse data and imperfect diagnostics. This approach allows for a robust assessment of diagnostic performance, particularly in resource-limited settings, ensuring that diagnostic tools provide meaningful data for surveillance and control efforts. These findings not only underscore the flexibility of Bayesian methods but also emphasize the importance of adhering to guidelines like STARD-BLCM to ensure transparency and rigor in diagnostic evaluations. Building upon the insights gained from test evaluation, the thesis advances to prevalence estimation, a critical step in understanding disease burden. A novel contribution of the work is the development of a novel Bayesian model that incorporates sequential testing schemes, where confirmatory diagnostics are applied only to positive results from initial screening tests. This innovation captures practical diagnostic workflows often used in real-world settings but rarely addressed comprehensively in the literature. Using brucellosis in livestock in Iran as a case study, the research provides refined prevalence estimates and highlights the endemic nature of the disease. By accounting for the biases introduced by sequential testing, this work advances the methodological application of Bayesian models and offers actionable insights for targeted interventions and control strategies. The progression of the thesis naturally leads to the concept of freedom from infection, an essential goal in disease eradication and control efforts. The work delves into methodologies for demonstrating freedom from infection, emphasizing output-based frameworks that focus on achieving confidence in disease freedom rather than adhering to rigid input-driven standards. Through a review and application of modern approaches such as scenario tree analysis, Bayesian hierarchical models, and the STOC Free model, the thesis explores how probabilistic methods can enhance surveillance Se and Sp. By integrating longitudinal data, population risk factors, and diagnostic test characteristics, these methods provide a nuanced understanding of infection dynamics and inform cost-effective strategies for substantiating freedom from disease. This aspect of the thesis aligns closely with international public health and trade priorities, offering a framework for robust and efficient surveillance system design. Finally, the thesis expands its scope to include disease surveillance and early outbreak detection, showcasing the application of the Epidemic Volatility Index (EVI) as an innovative tool for real-time monitoring. Applied to respiratory infections in Greece, including COVID-19 and influenza cases, the research demonstrates the operational simplicity and computational efficiency of EVI in detecting potential epidemic waves. By providing timely alerts and actionable insights, this work underscores the importance of proactive surveillance systems that can mitigate the impacts of epidemics and optimize resource allocation. This component highlights the broader applicability of the thesis's methodologies beyond zoonotic diseases, extending their relevance to other infectious and even non-infectious public health threats. In conclusion, the thesis illustrates how Bayesian methodologies and advanced algorithms can serve as a unifying framework for addressing diverse challenges in zoonotic disease management. From improving diagnostic accuracy and refining prevalence estimates to substantiating disease freedom and developing early warning systems, this work provides a roadmap for integrating advanced statistical tools into public health practice. By emphasizing the interconnectedness of human, animal, and environmental health within the One Health framework, the research underscores the necessity of multidisciplinary approaches in tackling zoonotic disease threats. Ultimately, the thesis offers not only a set of methodological advancements but also practical applications that contribute to the global effort to enhance public health resilience and readiness in the face of emerging and re-emerging zoonotic diseases.

Degree

thesis:*
Grantor dc:publisher
University of Thessaly (UTH)
Year dc:date
2024

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Meletis, Eleftherios
  • Μελέτης, Ελευθέριος

Subjects

dc:subject × 17

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/58083

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Meletis, Eleftherios; Μελέτης, Ελευθέριος. Application of Bayesian methods and advanced algorithms in the measurement of zoonotic disease occurrence. University of Thessaly (UTH), 2024. http://hdl.handle.net/10442/hedi/58083