Quantifying the information in noisy epidemic curves.

Area of Science:

• Epidemiology
• Mathematical Biology
• Public Health

Background:

• Accurate estimation of transmissible disease dynamics is crucial for public health policy.
• Surveillance data, commonly used for inference, suffers from inherent reporting errors and delays.
• These limitations critically impact the reliability of epidemiological parameter estimates.

Purpose of the Study:

• To develop an analytical framework for quantifying uncertainty in disease dynamics estimation.
• To create a metric for ranking the informativeness of surveillance data.
• To evaluate the reliability of case and death curves for inferring transmissibility.

Main Methods:

• Developed an analytical framework to quantify uncertainty from under-reporting and reporting delays.
• Introduced a metric to rank surveillance data informativeness.
• Applied the framework and metric to epidemic case and death time series data.

Main Results:

• Quantified the uncertainty introduced by under-reporting and reporting delays in disease surveillance.
• Found that death curves are not universally more reliable than case curves for inferring transmissibility.
• Demonstrated how surveillance limitations reduce actionable information on pathogen spread.

Conclusions:

• The reliability of epidemiological inferences is significantly affected by surveillance data quality.
• A nuanced approach is needed when selecting data sources for estimating disease transmission dynamics.
• The developed framework provides a quantitative basis for understanding and mitigating surveillance-induced uncertainty.