There are two distinct mechanisms for modelling and, more broadly, for science to directly inform policy in a pandemic or similar emergency:

(i) close and direct collaboration, likely led by government or public health institutions and likely focused on building tools to answer specific research-to-policy questions, and

(ii) consensus building at a wider scale with a broader community of academics including modellers and others.

In the close collaboration model for knowledge translation, public health epidemiologists, scientists, and modellers work side-by-side with academic partners either in virtual or physical spaces (or a combination of the two) to ensure seamless flow of public health needs, data, innovations, and modelling solutions. This creates outputs of great direct use for both public health and academia and also provides impetus for innovations in modelling.

In the consensus building model for knowledge translation, modellers develop multiple models and viewpoints in response to emerging threats, current focal questions and requests from decision-makers. This is a key part of the scientific and analytic process. For their part, researchers need to recognize that policymakers need a platform for timely decision-making based on a synthesis of models that enjoys consensus support within the academic community. Such a synthesis is not the endpoint of inquiry: rather it is a consensus view at a point in time that is both scientifically valid and “serviceable” for the purposes of informing policymaking (Jasanoff 2015). [1]

This means that the process for achieving this consensus needs to be inclusive of a range of modelling centres and networks. In contrast to close and direct collaboration, this approach needs to be sufficiently independent of government that it has credibility within the scientific community and with the public. However, scientists should expect that ongoing conversations with policymakers will inform and enrich the boundary between science and policy. Policymakers, on their part, need to recognize that there will be considerable uncertainty, particularly on short time frames, and differences of opinion and dissent will occur as the science continues to develop. However, they should be able to expect a synthesis of advice that flows from models, relevant to a particular question. The network we envision will support these activities.

More broadly, models play a role in building and communicating scientific consensus and “serviceable truths”, because of modelling’s ability to synthesize diverse data, explore the consequences of assumptions, quantitatively test assumptions against data, and incorporate information as it emerges. In this context, the modelling and scientific communities can support decision-makers by coming together to create platforms for decision-making, rapidly assessing these through formal or informal peer review and robust debate, and by coming to agreed conclusions that are informed by this evidence and transparent process.

Workshops and training are essential ingredients for knowledge translation in EIDM. They operate in both directions: training public health epidemiologists to build knowledge of modelling and training modellers to build knowledge of epidemiology and public health. This needs to be ongoing and provide core training for HQP. Activities include development of simulated pandemic exercises to underpin greater pandemic preparedness and mutual understanding of modelling for public health in day-to-day and emergency outbreak situations. They also need to include development of skills in communication to the public for both academics and public health professionals.

There are some essential organizational ingredients that are predicated by the above framework for knowledge translation to ensure long-term success in translating EIDM into public health gains.

• Public health and researchers must collaborate to develop a secretariat function that facilitates “evergreen collaboration” with academia (e.g., regular communication activities) and to ensure decision-makers obtain clear and consensus messages from the modelling community.

• Productive interactions with the EIDM community require that there are people embedded in public health and government who understand and appreciate the strengths and limitations of modelling.

• Federal public health institutions need to provide the opportunity for links with the global community in public health and international benchmarking of methods.