Johns Hopkins AI Tool Enhances Predictive Accuracy for Infectious Diseases

Researchers at Johns Hopkins University have developed an advanced artificial intelligence (AI) model that significantly enhances the predictive accuracy of infectious disease outbreaks, including COVID-19 and bird flu. This innovative tool, detailed in a study published in the journal *Nature Computational Science* on June 9, 2025, aims to improve public health responses by forecasting disease spread more effectively than existing methods utilized by the Centers for Disease Control and Prevention (CDC).

The AI model, led by Dr. Lauren Gardner, a professor at the Johns Hopkins Whiting School of Engineering and director of the Center for Systems Science and Engineering, utilizes large language modeling techniques akin to those found in popular AI applications like ChatGPT. This enables the model to predict epidemiological trends and hospitalization rates up to three weeks in advance. "We know from COVID-19 that we need better tools so that we can inform more effective policies," Dr. Gardner stated. "There will be another pandemic, and these types of frameworks will be crucial for supporting public health response."

Historically, infectious disease modeling has faced challenges due to the complexity and variability of factors influencing disease spread. Dr. Gardner noted that while traditional models performed adequately during stable conditions, they struggled to account for sudden changes, such as new viral variants or shifts in public health policies. The new AI tool addresses these limitations by integrating diverse data sources, including demographic information, public health interventions, and genomic surveillance data.

"Traditionally, we use the past to predict the future, but that doesn’t give the model sufficient information to understand and predict what’s happening," explained Dr. Hao “Frank” Yang, assistant professor of Civil and Systems Engineering at Johns Hopkins University. "Instead, this framework uses new types of real-time information."

In evaluating the model's performance, the research team retroactively tested its predictions against actual COVID-19 case data across all U.S. states over 19 months. The AI model outperformed existing predictive frameworks, including the CDC's CovidHub weekly forecasts for COVID-related hospital admissions. This marks a significant advancement in the ability to forecast public health crises effectively.

The implications of this research extend beyond COVID-19. The model is designed for flexibility and can be adapted to monitor various infectious diseases, such as monkeypox and the respiratory syncytial virus (RSV). Dr. Gardner emphasized that understanding the drivers of infection surges and integrating those insights into predictive models is crucial for enhancing public health responses.

Looking ahead, the research team plans to further explore how large language models can be utilized to comprehend and predict individual health-related decision-making. This could aid public officials in crafting policies that promote safer health practices and more effective disease management strategies.

This innovative application of AI in public health forecasting represents a pivotal step toward better preparedness for future pandemics. By harnessing advanced computational techniques, researchers at Johns Hopkins University are not only improving predictive accuracy but also paving the way for more informed public health interventions that could save lives during infectious disease outbreaks.