Vaccine and Infectious Disease Division

Optimizing vaccine allocation during flu epidemics

In pandemic outbreaks of influenza, the epidemic can spread rapidly before large-scale vaccine production can catch up.  This was particularly notable in the 2009 H1N1 epidemic, where an initial outbreak occurred in April, but vaccine was not available until after a secondary autumn outbreak was underway.  With a limited supply of vaccine, it is essential to determine the optimal vaccination allocation pattern to minimize the outbreak and to best protect the most vulnerable populations.  Complicating the strategy, it might be more important to vaccinate high-transmission groups such as school children early on during an outbreak, while later in an epidemic it might be more effective to vaccinate groups at high risk for complications from the disease, such as the elderly, pregnant and immunocompromised.

To evaluate these hypotheses, UW Applied Mathematics doctoral student Laura Matrajt and her advisor, VIDD member Ira Longini, developed a deterministic model of flu vaccination.  They considered two population structures: one from a developed country where children account for 24% of the population, and another from a less-developed country where children make up 55% of the population.  In both groups, they considered vaccine availability and initial vaccination time, adjusting the model under each scenario to minimize mortality and hospitalizations.

Their model demonstrates that the optimal vaccination strategy varies depending on the population and timing of vaccination.  They found that in the optimal vaccination strategy, a switch occurs shortly before the epidemic peak, at which time it becomes more efficacious to vaccinate high-risk individuals than high-transmission individuals.  A challenge remains in determining the current stage of an epidemic outbreak, to decide whether to target high-transmission or high-risk individuals for vaccination.

Matrajt L, Longini IM Jr.  Optimizing vaccine allocation at different points in time during an epidemic.  PLoS One. 2010 Nov 11;5(11):e13767.