Flu Prediction Market, 2004-05

The home site for the 2005-06 edition of the Influenza Prediction market has moved to http://fluprediction.uiowa.edu Please change your bookmarks.

Welcome to the 2004-2005 Influenza Prediction Market homepage.

The 2004-2005 Influenza Prediction Market opened for trading on Monday, September 20, 2004.

Influenza is a major cause of morbidity and mortality throughout the world. Unlike many viral diseases, a vaccine and effective prophylactic medications are available. However, the effectiveness of these measures depends on timing. Antiviral compounds must be given in the first 48 hours of illness, and the vaccine does not reach peak protective effect until two weeks after administration.

Influenza occurs on a predictable seasonal basis during the winter months and there is evidence that it has done so for hundreds of years. Despite the seasonal predictability of influenza, considerable variability exists regarding the extent of disease activity as well as the timing and geographical location of influenza outbreaks. Thus, accurate forecasting of influenza activity even 1-2 weeks in advance would assist efforts to modify the severity of influenza at both individual and community levels.

Information about influenza activity is diverse and widely distributed. Different health care professionals have different information regarding influenza activity. For example, a school nurse might observe higher-than-expected absence rates, an emergency room physician might note that more patients with influenza-like illnesses are seeking treatment, and a clinical microbiologist might observe a sudden increase in respiratory cultures positive for influenza virus. This information could be quite helpful in predicting future influenza activity if it could be aggregated and analyzed efficiently. Because this information is disparate, standard research and statistical methods have not proven to be effective. Thus, the medical community does not have access to accurate influenza forecasting. The Influenza Prediction Markets described on these pages is an attempt to satisfy the need for accurate information regarding future influenza activity.

Experimental markets have been used by the Iowa Electronic Markets to predict unusual commodities ranging from election results to movie box office receipts. Since 1988 the Iowa Electronic Markets have developed a forecasting record substantially superior to alternative mechanisms. We propose that markets for infectious diseases may be useful for predicting infectious disease activity quickly, accurately, and inexpensively by aggregating the expert opinion of health care professionals.

The Influenza Market is a futures market and operates very much like the commodity markets in Chicago and New York, though the "contracts" traded are claims to future payments that depend on the timing and extent of influenza activity. For example, a "RED_52" contract might promise to pay one dollar if the CDC determines that influenza was widespread in Iowa during the last week of 2004. Participating medical professionals will buy these contracts when they see evidence suggesting to them that influenza will be widespread in Iowa for that particular week, and they will sell if the evidence is not there. The efforts to buy and sell push the price of the RED_52 contract up and down respectively. Observers of the market can then use this price as a prediction of the level of flu activity in Iowa during the last week of 2004.

 Interested in joining the market? — Participation in the Influenza Prediction Markets is by invitation only. Invitations will be extended to medical professionals with interest or expertise in influenza. We plan to enroll about 100 traders throughout the state of Iowa. We are interested in recruiting: primary care nurses and physicians, pharmacists, infection control practitioners, clinical microbiologist, lab technicians, and epidemiologists.

We plan to enroll about 100 traders throughout the state of Iowa. If you think you qualify and are interested in joining, please contact Philip Polgreen, MD at the University of Iowa, Carver College of Medicine, Division of Infectious Diseases, philip-polgreen@uiowa.edu. In that email, provide the following information:

  • Your name
  • Email address
  • Postal address
  • Phone number
  • Identification number (any unique 4-9 digit identifying number you that you will remember)
  • Professional position
  • Brief explanation of your expertise with influenza
  • How you learned about the market

 Contact Information — The Flu Prediction Market is a collaborative research effort involving members of the Carver College of Medicine and of the Tippie College of Business at the University of Iowa. Market Managers can be reached by email at philip-polgreen@uiowa.edu or forrest-nelson@uiowa.edu