Collabathon on Estimating Real-Time Epidemic Growth Rates

In September, Wake Forest University School of Medicine researchers Michael DeWitt and Brinkley Bellotti from the Section on Infectious Diseases attended the 2024 Epistorm Rt-Estimate Collabathon in Boston hosted by Northeastern University and the Network Science Institute. A collabathon (combination of “collaboration” and “hackathon”) is a multi-day event where practitioners and researchers come together to build software and form collaborations around one main idea. This collabathon included experts from public health agencies, industry, academia, and NGOs brought together to address a key challenge in public health: estimating real-time epidemic growth rates. Supported by the Centers for Disease Control and Prevention CFA Insight Net initiative, the collabathon focused on improving tools for estimating the real-time reproduction number: a key epidemiologic parameter that drives infectious disease control and prevention measures. Some of the major outputs of this workshop included new frameworks for testing modelling methods as well as new software packages to improve estimations.

“I worked primarily in the Julia programming language to implement a package to handle primary censoring and truncation,” said Michael DeWitt. “This phenomenon occurs because we don’t know the precise moment when someone is infected and similarly delays in our detection lead to both censoring and truncation. Outbreak and forecasts that don’t take this phenomenon into account will be biased. By developing modular software in Julia, we can allow researchers and infectious disease modelers better, plug and play tools to forecast outbreaks.”

“Though many methods and tools to estimate Rt exist, techniques to evaluate model performance are still needed. Because the goal of real-time epidemic growth rate epidemics is prediction, the team I worked with began developing methods to validate Rt estimation methods with short-term prediction performance,” said Brinkley Bellotti.

Additional package development included summRt to help provide summary statistics for Rt estimates and RtEval to help evaluate the performance of Rt estimation methods.

This work continues to underline Wake Forest University’s School of Medicine’s commitment to research and innovation in microbial ecology and emerging infectious diseases.

@online{dewitt2024,
  author = {DeWitt, Michael and Bellotti, Brinkley},
  title = {Collabathon on {Estimating} {Real-Time} {Epidemic} {Growth}
    {Rates}},
  date = {2024-10-28},
  url = {https://wakeforestid.com/posts/2024-10-28-epistorm-rt-collabathon/},
  doi = {10.59350/mcqh1-nvd03},
  langid = {en}
}