The role of mathematical modelling

Over the past five years, efforts to mathematically model the risks, costs, and benefits of future strategies for managing polio yielded several key insights. Although models cannot predict the future, they provide glimpses of the many possible scenarios and thereby help define the most effective management strategies for the long-term.

Modelling the numerous future policy options related to routine vaccination, supplemental vaccination, surveillance, outbreak response, vaccine stockpile, containment, and other issues clearly demonstrates the need to recognize the differences that exist in national values, preferences, and risks, as well as the interactions among the policy options. Efforts to quantify the risks show that continued use of oral poliovirus vaccine (OPV) for routine vaccination after successful interruption of wild poliovirus transmission globally will mean continued costs and cases associated with vaccine-associated paralytic polio (VAPP) and outbreaks of circulating vaccine-derived polioviruses (cVDPVs).

Even stopping OPV will not immediately mean the immediate absence of paralytic polio, because the world faces a relatively high probability of experiencing one or more outbreaks from a cVDPV in the first five years after OPV cessation (decreasing probability for each year after OPV cessation). Although the risks of an outbreak in any individual country may seem small and it is not possible to predict where or when such future events will occur, the modelling results suggest that:

1. high quality surveillance will need to continue globally
2. response teams will need to be prepared to aggressively contain any outbreaks detected
3. a stockpile of vaccine will be needed for outbreak response.

The modelling also suggests that cessation of OPV should be coordinated globally, and that countries should understand the increased costs and risks they might incur by stopping either earlier or later than the rest of the world.

With respect to outbreak response, a dynamic model based on the epidemiological and virological data from current and historical outbreaks identified conditions that tend to lead to larger occurrences, notably the time to detect and respond to the outbreak. Based on this modelling research in 2005, the Global Polio Eradication Initiative implemented strategies to reduce its outbreak response times. Modelling also played a key role in 2007 in helping policy makers understand the cost and case implications of abandoning the goal of eradication to switch to control.

Although models provide the opportunity to explore the potential uncertain futures and our options, they are limited by the quality of the information going into the models and by the amount of complexity that they can capture. However, modelling represents a critical tool that helps stakeholders understand the trade-offs that come with different choices in the context of an uncertain world.



Outbreak size as a function of response delay and coverage in a low-income country with 10 million people, no routine immunization for five years after OPV cessation, acute flaccid paralysis surveillance, and low-medium R0 (=10). All response scenarios target children less than five years of age with mOPV and occur at 30-day intervals with the second and third rounds achieving 90% coverage.