Comparing methods to quantify experimental transmission of infectious agents

Area of Science:

• Epidemiology
• Mathematical Biology
• Infectious Disease Modeling

Background:

• Quantifying infectious agent transmission is crucial for understanding epidemics.
• The transient-state (TS) algorithm, based on the stochastic SIR model, offers time-dependent epidemic probability distributions.
• Numerical limitations restrict the TS algorithm's application to small populations.

Purpose of the Study:

• To investigate the errors associated with using the final-size (FS) algorithm when the final-size state is not reached.
• To introduce and evaluate a new method, the MaxDiff test, for detecting transmission differences between groups.
• To compare the statistical power of FS, TS, and MaxDiff tests in identifying transmission variations.

Main Methods:

• Analysis of errors introduced by applying the time-independent FS algorithm in non-final-size scenarios.
• Development and application of the MaxDiff test for comparing transmission rates between treatment groups.
• Comparative power analysis of FS, TS, and MaxDiff tests using simulated or experimental epidemic data.

Main Results:

• Methods based on the FS algorithm underestimate R(0) and exhibit biased hypothesis testing outcomes.
• The FS algorithm is less powerful, particularly when transmission rates in both control and treatment groups exceed one.
• The TS and MaxDiff tests demonstrate superior power in detecting differences in transmission between groups.

Conclusions:

• The FS algorithm introduces significant errors and reduced statistical power when applied prematurely in epidemic modeling.
• The novel MaxDiff test, alongside the TS algorithm, provides a more powerful approach for analyzing transmission dynamics and inter-group differences.
• Accurate quantification of infectious disease transmission requires careful consideration of the chosen algorithmic approach and epidemic stage.