Pitt Researchers Propose New Model To Design Better Flu Shots

The flu shot, typically the first line of defense against seasonal influenza, could better treat the U.S. population, thanks to University of Pittsburgh researchers. New research that focuses on the composition and timing of the shot design was published in the September-October issue of Operations Research by Pitt Swanson School of Engineering faculty members Oleg Prokopyev, an assistant professor, and Professor Andrew Schaefer, both in the Department of Industrial Engineering, and coauthors Osman Ozaltin and Mark Roberts, professor and chair in Pitt's Department of Health Policy and Management. Ozaltin, who is now an assistant professor of engineering at the University of Waterloo in Ontario, did his research for the study as a Pitt graduate student in the Swanson School; he earned his Pitt PhD degree in industrial engineering earlier this year.

The exact composition of the flu shot is decided every year by the Food and Drug Administration (FDA), and the decision is complicated.

"The flu's high rate of transmission requires frequent changes to the shot," said Prokopyev. "Different strains can also cocirculate in one season, which gives us another challenge for figuring out the composition."

The Pitt researchers used powerful optimization methods from engineering to examine whether they could improve the yearly decisions made regarding what strains of influenza should be included in the current year's vaccine. The strains of flu that will be most likely to appear in the regular flu season are not known with certainty, but waiting longer to finalize the composition of the vaccine and observing what strains are occurring in other parts of the world improves the accuracy of the selection. However, the longer the FDA waits to make the decision, the more likely it is that there will be insufficient vaccine produced by the start of flu season. The model developed by the Pitt researchers balances these two important characteristics of the flu selection decision and integrates the composition and timing decisions of the flu shot design.

The model allows examination of the effect of many changes to the design and production of the vaccine, such as how many strains to include in the shot, when to make the final decision, how many times the FDA should meet to re-examine the current information concerning strains in other parts of the world, and the potential benefits from improved production methods.

"With this model, several policy questions can be addressed," said Schaefer. "For example, incorporating more than three strains might increase the societal benefit substantially, particularly under more severe flu seasons."