UI Researchers Receive Grant To Expand Influenza Prediction Market
The Iowa Electronic Markets seeks to target, in advance, which influenza strains may spread in the United States and where outbreaks will occur. This may allow public health officials, vaccine producers and health providers to better anticipate where flu activity will be most intense, and prepare for and respond to outbreaks with greater effectiveness.
The venture uses a small-scale financial market to aggregate the expert opinions of health care professionals. Epidemiologists, clinicians, pharmacists, nurses, and clinical microbiologists will buy and sell color-coded contracts representing levels of influenza activity, from yellow (no activity) to red (widespread activity), in the coming weeks. The probability of influenza spreading or declining will be reflected in the prices at which the contracts trade.
The market began as a pilot project in February 2004 with a focus on influenza activity in Iowa. Forrest Nelson, University of Iowa economics professor, said the RWJF grant support will be used to expand the market nationwide over the next five years. Co-investigators on the influenza prediction market are George Neumann, professor of economics, and Phil Polgreen M.D., infectious disease physician in the department of internal medicine.
"The market can capture information that standard influenza surveillance systems might not identify," said Polgreen. "A pediatric nurse might see an increase in the number of children with flu-like illnesses, or a pharmacist might notice an increase in the number of prescriptions for antiviral medications being filled. This kind of information is helpful in tracking and predicting influenza activity. We believe that prediction markets provide an efficient way to aggregate information that might be ignored by traditional disease surveillance methods."
Influenza causes substantial morbidity and mortality. In the U.S. alone, it causes an estimated 36,000 deaths and 120,000 hospitalizations annually. Influenza recurs each year in regular cycles, but the geographic location, timing, rates of increase and decline, duration and size of each outbreak varies.
Despite the dangers of influenza and the potential benefit from advance warning, forecasts of activity are not available. This expansion project seeks to:
- develop and refine short-term prediction markets to forecast influenza activity up to eight weeks in advance on a local, regional and national scale, and also predict dominant strains in the annual influenza cycle;
- study the effectiveness and efficiency of different prediction market designs; and
- provide influenza forecasts to health care providers and public health officials.
"This project presents us with the unique opportunity to test whether prediction markets can help public health officials, medical professionals, and the public prepare for and respond more effectively to influenza outbreaks," said Robert G. Hughes, chief learning officer at the Robert Wood Johnson Foundation. "This innovative method of information aggregation and forecasting may serve as a valuable tool that can complement traditional disease surveillance methods and guide decisions by public health leaders, not only in combating influenza but other potential health threats as well."
During the influenza market's pilot phase, it was used to predict activity exclusively within Iowa and only health care professionals took part as traders. According to Nelson, doctors, nurses, microbiologists, pharmacists, and others who have access to information that may help predict the spread of the virus will be invited to participate as the project now expands throughout the country.
According to Nelson, the 2005-06 market will open for trading on Sept. 26. A portion of the RWJF grant funding will be used to seed trading accounts, so that participating health care workers will not have their own money at risk. They will, however, receive the "profits" from their trading in the form of educational grants, thus motivating their active participation in the market and willingness to share information regarding influenza activity.
The influenza market is an outgrowth of the Iowa Electronic Market (IEM) project in the Tippie College of Business. That parent project, best known for its success in predicting outcomes ranging from political elections to movie box office receipts, has operated since 1988 and has established an impressive prediction record, substantially superior to alternative mechanisms such as opinion polls. Neumann and Nelson were two of the originators of the IEM project. They teamed with Polgreen to expand the prediction market technology to the field of medicine.
The Robert Wood Johnson Foundation is the nation's largest philanthropy devoted exclusively to improving health and health care for all Americans. Funding was provided through the Foundation's Pioneer Portfolio, which seeks to identify innovations that can lead to fundamental breakthroughs in health and health care.
Contact: Forrest Nelson, Iowa Electronic Markets, 319-335-0854