Leveraging forecasts to optimise decision making

The COVID-19 pandemic underscored the need for epidemic forecasts to guide policy decisions on resource allocation, public health measures, and risk communication.

However, deriving outputs from epidemic forecasts that provide answers to the questions that public health practitioners and policy makers actually have is not commonly done.

Additionally, historical forecasts are frequently not openly accessible for forecast performance to be evaluated, which makes it harder for policymakers looking to gauge the reliability and trustworthiness of a forecast and determine the extent to which its predictions should influence subsequent public-health decisions.

Our proposal aims to close the interpretability gap that precludes optimal use of epidemic forecasts.

We will first conduct a series of structured interviews with public health practitioners and external stakeholders to better define unmet needs around epidemic forecasting and identify common policy-relevant needs.

Secondly, we will develop a software package and web-based tool aimed at enabling policymakers, especially in contexts where technical capacity is limited, to better understand, interpret and action epidemic forecasts.

Our vision is to develop a computational tool to retrospectively assess forecast accuracy, defining accuracy based on policy-relevant derived statistics identified from our initial interviews, and transform them into interpretable and actionable information.