Methods for inference on transmission in seroprevalence data for multiple infections

Area of Science:

• Epidemiology
• Infectious Disease Research
• Biostatistics

Background:

• Identifying risk factors for novel infectious diseases is challenging.
• Utilizing data on seroprevalence of multiple infections in populations is valuable.
• Using covariate infections in analyses can be problematic if not independent risk factors.

Purpose of the Study:

• To propose a strategy for analyzing infectious disease risk factors when unknowns exist.
• To leverage parallels with known infections (covariate infections) to understand new agents.
• To develop hypotheses for disease transmission using the method of analogy.

Main Methods:

• Avoid adjusting for covariate infections when estimating effects of measured risk factors to prevent overadjustment.
• Use the estimated effect of covariate infections after controlling for measured factors as an indicator of unmeasured shared risk factors.
• Employ methods for repeated measures of categorical variables to infer shared mechanisms when shared measured risk factors are present.

Main Results:

• The proposed analytic strategy helps identify the presence of unmeasured shared risk factors.
• It allows for inference about shared mechanisms of action when shared measured risk factors exist.
• This approach facilitates hypothesis development for understanding transmission of new infectious agents.

Conclusions:

• The method of analogy, by studying parallels with known infections, aids in understanding novel infectious diseases.
• The recommended analytic strategy provides a robust framework for investigating infectious disease risk factors.
• This approach is particularly useful for building understanding of disease transmission and generating research hypotheses.