2024/12/07 –, Main Stream
言語: English
Infectious diseases are the third leading cause of death worldwide, claiming more than 13 million lives every year. While there are many interventions available for managing epidemics – including vaccines, tests and treatments – the best approach is not always clear. There are complex dynamics at play, with many interacting variables, and the optimal response may vary significantly depending on outbreak severity, funding available, and even sociopolitical context.
Computational modelling is a highly effective way to examine the potential impact of different public health interventions. However, the technical difficulty of building fit-for-purpose models that can simulate these impacts and produce quantitative optimisations is a barrier to widespread use.
Atomica is a Python-based simulation engine for compartmental modelling. It provides an easy-to-use yet highly configurable way to build disease models, with built-in support for interventions and optimisation with budget constraints. By leveraging population, transmission, and public health intervention data, as well as conceptual knowledge of how people progress through stages of disease, we can forecast the potential consequences of different public health strategies in specific countries or settings. With this insight, we can help governments to make the best possible decisions on what to prioritise, saving lives in the process.
Whether you’re a beginner or an expert Pythonista, this talk will equip you with the tools to simulate a real-world epidemic and optimise funding distribution for maximum impact.
Description for reviewers
In this talk, I will introduce attendees to a simple SIR disease model ( [Susceptible] --> [Infected] --> [Recovered] ). I will walk them through how to use the Atomica package to run a simulation of this disease model (or any infectious epidemic, as the steps are the same), using the in-built optimisation algorithm to generate a bar plot of the recommended funding allocations for a given budget.
Since Atomica has so much in-built functionality for modelling these systems, the walkthrough is simple enough that the topic of the presentation can easily be covered in 15 minutes.
Outline
Introduction (2 min):
• Global impact of infectious diseases
• Basic infectious disease dynamics ( [Susceptible] --> [Infected] --> [Recovered] )
Using Atomica to model a simple epidemic (8-9 mins total):
• 1. SYSTEM FRAMEWORK and DATA: Implementing the disease dynamics structure and entering its data values (2-3 mins)
• 2. SIMULATION: Running the simulation with run_sim() (1 min)
• 3. CALIBRATION: Adjusting parameter values with calibrate() (2 min)
• 4. INTERVENTIONS: Adding simple public health interventions (3 min)
Modelling results (3 mins):
• Outputs: Bar plot of optimal funding allocations. How many lives saved.
• Modelling results from actual tuberculosis work
Conclusion (1-2 min):
• Python’s usefulness for simulating epidemics and predicting outcomes
• Link resources to learn more about Atomica
Abstract Summary
Infectious diseases are a major global threat, causing millions of deaths every year. However, managing an epidemic is complex due to the many interacting variables and the wide range of contexts they occur in.
Atomica – a Python-based simulation engine for compartmental modelling – is a powerful tool for simulating epidemics and forecasting the impacts of public health interventions. The insight this tool provides can help governments make life-saving descisions. In this talk, you will learn how to model an epidemic and optimise funding distribution for maximum impact.
¡Hola! Soy Eloisa, una modeladora apasionada por la programación.
Estudié física en la universidad de Sydney, con enfoque hacia la programación y biotecnología. Ahí investigué cómo hacer nanorobots médicos para el cerebro, desarrollando algoritmos ML para analizar mapas cerebrales, y creando puertas lógicas hechas de moléculas en un laboratorio.
Llevo desarrollando y utilizando modelos compartimentales desde 2021, primero en el ámbito de la salud mental y ahora en el contexto de las epidemias. Mi trabajo actual en el Instituto Burnet se centra en evaluar el futuro impacto de intervenciones sanitarias y nuevos medicamentos, y a veces también traduzco resultados e informes al castellano :)