What if we could forecast the flu like we forecast the weather?


The flu affects millions of people every year. Advanced warning of a local outbreak could help people prepare. / Image: Adobe Stock

The local weather forecast helps us plan and prepare for severe storms. What if we could forecast disease outbreaks in the same way? That would allow government officials, health care providers, business leaders and concerned patients to make life-saving decisions.

How to predict flu outbreaks

The Optum flu forecast is a data-driven early-warning system for influenza and related illnesses. It grew out of a relationship between Optum and Carnegie Mellon University (CMU), which is renowned for its computer science and artificial intelligence research. Teams at Optum applied their expertise and data assets to expand the technology’s scope and capabilities. The Optum flu forecast was born.

The data used for the Optum flu forecast comes from extensive array of sources, including:

  • De-identified medical and pharmacy claims records related to flu diagnosis
  • Health data and clinical reports from the Centers for Disease Control and Prevention
  • Publicly available social data
  • Flu-related web search queries

“Any one of these data sources on their own would not necessarily be that great, but all of them together are quite good and quite robust,” notes Professor Roni Rosenfeld of CMU. He said the data from Optum is some of the best he has worked with. And unlike some other sources, it’s geographically specific, not aggregated at a state level. This can help predict flu outbreaks down to a metro and county level, two to four weeks before they occur.


Source: Centers for Disease Control and Prevention.

Preventing the flu saves lives

Each year, tens of thousands of people die from influenza. Hundreds of thousands are hospitalized. Millions more are sick at home. Flu season can last many months. Prevention efforts can be difficult to sustain long term. But if you know a storm is coming, you can batten down the hatches and prepare for it.

It’s estimated that thousands of deaths (and millions of illnesses) were prevented by the flu vaccine last year. But it takes about two weeks after vaccination for antibodies to develop in the body and protect against the flu.

Communities can help

With more than two weeks’ notice of an outbreak in a specific area, people could be reminded to get a vaccine. Those most at risk could decide if they’re better off staying home.

Clinics could separate people with flu symptoms from others in waiting rooms, and provide face masks to minimize the risk of transmission. Health systems could reroute supplies, medication and staff to help protect their patients, members, employees and communities.

Employers could get the message out to their workers about the importance of a flu shot and proper hand hygiene. They could set up on-site vaccination clinics. They could allow more people to work from home, conduct virtual meetings, or find ways to reduce the number of employees in confined spaces.

The power of data and technology

Advances in data curation and emerging technologies won’t eliminate the flu, but they can improve our ability to prevent and combat it. The long-term vision is to make disease forecasting as accepted and useful as weather forecasting is today. Just as you can check the weather forecast to decide if you need an umbrella, one day you could check the flu forecast to decide how to prepare for that, too.

To read more about the Optum flu forecast, read the full article, “Turning the tide on flu: The promise of data-based forecasting.”

And to learn more about technology innovation at Optum, visit optum.com/technology.

Leave a Reply