Sizing Up Stockpiles of Children's Vaccines

A creative version of a classic engineering technique may improve decisions about building and using supplies of important pediatric vaccines, potentially leading to lower public health costs and healthier children.

The United States maintains a six-month supply of common pediatric vaccines to ensure protection from deadly diseases, such as the flu, polio, and diphtheria, despite interruptions in vaccine production. The stockpiles must be replenished as the vaccines are used or expire, and, because the manufacture of vaccines is a laborious and unreliable process, health officials must place orders for new vaccines up to a year in advance.

Researchers at the University of Illinois at Urbana-Champaign (UIUC) and the Rochester Institute of Technology (RIT) have developed a mathematical framework to better understand the implications of vaccine stockpile levels through evidence-based engineering principles. Industrial engineers Sheldon Jacobson of UIUC and Rubén Proaño of RIT, who specialize in operations research, and Janet Jokela, a specialist in public health and infectious diseases at UIUC, published this work in the online edition of the November 2010 Journal of Industrial and Management Optimization.

Deciding how many pediatric vaccine doses to order from year to year is no simple task. According to the researchers, "The decision must balance several objectives that sometimes conflict." These include: minimizing the impact of vaccine shortages, maintaining or increasing vaccine coverage, and minimizing vaccine costs (including costs from unused vaccine).

The number of doses to order also may also depend on the importance of the vaccine. Some vaccines are easier to obtain than others, some diseases are more contagious or more deadly than others, and society has higher immunity levels against some pediatric diseases than others.

The researchers' model for setting stockpile levels is a novel adaptation of the "utility maximization problem" (UMP). One everyday example of a UMP is the set of considerations involved in choosing a car--such as cost, performance, and gas mileage--that a buyer must weigh based on relative importance. The best car for a particular buyer will depend on his or her preferences.

"UMPs have long been used by engineers and businesspeople to optimize decisions," explains Jacobson. "What's unique about this work is the way we customized the UMP to take into account multiple objectives and criteria." The mathematical framework allows health officials to see the optimal stockpile levels for different initial conditions and preferences of public health officials.

"The framework developed by Jacobson and his collaborators begins a new and richer dialogue about vaccine stockpiling," says Russell Barton, NSF program director for Service Enterprise Systems. "The likely result will be better decisions on setting vaccine stockpiling policies. But the framework of multi-attribute utility theory has the potential to transform many processes for setting health-care policy. "

Through eight hypothetical scenarios, the researchers demonstrated how different approaches to managing stockpiles of six pediatric vaccines have different implications for public health. When initial conditions are poor, due to low stockpiles and/or low vaccine coverage, the scenarios revealed that the preferences of the public health decision-maker can significantly affect what the optimal stockpile size would be.

For example, if vaccine coverage is high but the stockpile inventory is low, a preference for minimizing the impact of vaccine shortages might focus vaccine resources on just a few important diseases. Under these same conditions, a decision-maker who also considers vaccine coverage and cost would focus resources on only the one most important disease.

"In a number of likely scenarios," Jacobson says, "our research indicates that one size does not fit all when it comes to the optimal size of the vaccine stockpiles." Depending on the relative importance of a disease, health officials may decide to maintain more or less than a six-month supply of vaccine.

According to Jacobson, the model also demonstrates that the vaccine stockpile could be used strategically to actively increase vaccine coverage or to protect society from disease outbreaks. He says, "At the beginning of an especially virulent disease outbreak, for example, it may even make sense to deplete the vaccine stockpile."

The researchers say that their approach could be modified to include other factors or conditions involved in decisions about vaccine stockpiles. One avenue to explore within the framework is the impact of a potential vaccine shortage in terms of the number of fatalities or QALYs (quality-adjusted life-years, which take into account both quantity and the quality of life generated by health care interventions). Another area of investigation, given the long-lasting effect of vaccination, is to extend the current model over multiple time periods.