Researchers are analysing various mathematical and computational simulation models to support evidence-based planning and policies during the COVID-19 pandemic.
Information about the different models will be used to inform end-users of the model objectives and policy issues, methods, assumptions and data in a clear and comparable way. Policymakers can then decide which models are fit for purpose on their specific policy questions.
The COVID-19 Multimodel comparison collaboration project (CMCC) is convened by Bill and Melinda Gates Foundation, the Centers for Disease Control and Prevention, the Department for International Development (UK), the Global Partnership for Sustainable Development Data, the international Support Decision Initiative (iDSI), the Norway Agency for Development Cooperation (NORAD), the Royal Thai Government, USAID, the World Bank and the World Health Organization (WHO).
The project comprises three groups that will interact with each other: an independent technical modelling group, a policy maker group, and a group of experts responsible for the existing major models being used, such as Imperial College London, Oxford University, the London School of Hygiene and Tropical Medicine and the Institute of Health Metrics and Evaluations in Washington University.
Policy experts will be consulted on how to best guide the selection and use of models, presentation of results, policy scenarios and outcomes that will be most relevant to LMICs to inform future model iterations.
Dr Itamar Megiddo, of the Department of Management, who is part of the initiative and is primarily involved in the technical modelling group, said: “There are different models out there that policymakers can use to decide the best approaches to dealing with the COVID-19 pandemic.
“Many LMICs were asking organisations like WHO which models they should be using to predict how COVID-19 might spread in their countries.
This collaboration has been put together to evaluate different models and help policymakers understand which approach is fits their purpose.
“Different countries will have different circumstances and thus the best intervention – such as lockdown, quarantine, track-and-trace – will be different.
“The accuracy of a model for specific countries depends on the data you can feed into it. Ideally you want data from the country you are modelling: population characteristics, health system characteristics, disease biology and so on.
“If you don’t have that data, you look for data from similar countries. If you don’t have that, you make an educated guess based on international data, and combining models.”
The technical group’s part of the project is divided into two steps and will possibly be extended further. The first step is to understand the different assumptions the different models are making and the different results they produce.
Phase two will see the development of data sets with common assumptions that all of these models can use. The researchers will then evaluate how different model assumptions lead to differing results and what the priorities are for data collection, and based on these analyses they will develop reports and recommendations.