Related Experiment Videos
Models for the statistical analysis of infectious disease data.
M Haber1, I M Longini, G A Cotsonis
1Department of Statistics and Biometry, Emory University, Atlanta, Georgia 30322.
Biometrics
|March 1, 1988
Summary
This study generalizes the Longini-Koopman model for infectious disease transmission by incorporating risk factors at household and community levels. It presents two model types for analyzing epidemic data, enhancing disease spread understanding.
Area of Science:
- Epidemiology
- Biostatistics
- Mathematical Modeling
Background:
- The Longini-Koopman model (1982) established a framework for infectious disease transmission probabilities within households and communities.
- Existing models often lack the granularity to account for varying risk factor levels influencing disease spread.
Purpose of the Study:
- To generalize the Longini-Koopman model by integrating household and community-level risk factors.
- To develop and analyze two distinct modeling approaches for infectious disease transmission data.
Main Methods:
- Developed generalized models accommodating different transmission probabilities based on risk factor levels.
- Considered two model types: (i) household data models and (ii) individual data models.
- Employed maximum likelihood estimation for both model types and weighted least squares for log-linear represented household models.
Main Results:
- Demonstrated that household data models are special cases of individual data models, analyzable as log-linear models.
- Successfully applied the generalized models to real-world influenza epidemic data from Tecumseh, Michigan, and Seattle, Washington.
- Validated the utility of both maximum likelihood and weighted least squares methods in analyzing the proposed models.
Conclusions:
- The generalized Longini-Koopman model effectively incorporates risk factors to provide a more nuanced understanding of infectious disease transmission dynamics.
- The presented methodologies offer robust statistical frameworks for analyzing epidemic data, enhancing public health preparedness and response strategies.