Influenza has caused seasonal epidemics for centuries and, on rare occasions, devastating pandemics. Tracking its historical impact helps researchers understand transmission patterns, intervention strategies, vaccine effectiveness, and the potential severity of future outbreaks. Several organizations, including the CDC and WHO, maintain detailed records of influenza occurrence and impact.

CDC Data

The U.S. Centers for Disease Control and Prevention (CDC) has tracked influenza activity since the early 20th century. The CDC FluView surveillance system compiles weekly data on influenza-like illness (ILI), lab-confirmed cases, hospitalizations, and deaths across the United States. CDC's publicly available databases enable longitudinal studies of flu burden, vaccine coverage, and viral strain prevalence.

WHO Data

The World Health Organization (WHO) coordinates the Global Influenza Surveillance and Response System (GISRS), established in 1952. GISRS collects flu data from over 100 countries through National Influenza Centers and collaborating labs. WHO uses this data to: (1) monitor influenza strains worldwide; (2) make biannual vaccine composition recommendations; and (3) track the global impact of pandemics like the 1918 Spanish Flu, 1957 Asian Flu, 1968 Hong Kong Flu, and 2009 H1N1.

Other's Data

Additional influenza data come from academic research groups, nonprofit organizations, and collaborations:

• FluNet (WHO's public-facing data portal)

• GISAID (Global Initiative on Sharing All Influenza Data)

• Project Tycho (University of Pittsburgh)

• IHME (Institute for Health Metrics and Evaluation)

Mathematical and statistical models are essential tools for understanding the spread, evolution, and control of influenza. These models help researchers and policymakers estimate key parameters such as transmission rates, predict epidemic curves, evaluate vaccine effectiveness, and simulate intervention strategies. Modeling approaches range from mechanistic systems of differential equations to data-driven machine learning and hybrid frameworks.

ODE-based models

Ordinary Differential Equation (ODE) models are a cornerstone of influenza modeling. They describe how individuals move between disease states (e.g., Susceptible–Infected–Recovered, or SIR models) over time. These models can incorporate age distribution in the population, vaccination, immunity waning, seasonality, and multiple influenza strains.

Statistical models

Statistical models focus on inferring parameters and making predictions from data. Time series models (e.g., ARIMA), regression-based approaches, and hierarchical Bayesian models are commonly used to estimate influenza incidence, detect season onset, or evaluate the impact of vaccination or other intervention programs. 

Other Models

Agent-based models (ABMs) simulate interactions among individual agents (e.g., people, households, schools). These models are similar to SIR models except they represent individuals as defined entities, wheras SIR models represent populations.

Network models model transmission over contact networks. This type of model is ideal for studying targeted interventions, like school closures or vaccination of specific subgroups.

Machine learning models are data-driven approaches (e.g., random forests, neural networks) that are increasingly used for influenza forecasting and outbreak detection. These models attempt to extract information from the data without needing a rational basis in the underlying biology.

Hybrid models (e.g., those that integrate ODE-based structure with statistical or machine learning components) attempt to incorporate flexible, data-informed analysis with mechanistic grounding.