At a glance
We explain the methods behind CFA’s outbreak scenario assessments. Subject-matter experts consider available evidence to assess what could occur in the coming weeks to months of an outbreak. Outbreak scenario assessments inform public health professionals and partners engaged in planning and executing emergency response activities.
Overview
CDC's Center for Forecasting and Outbreak Analytics (CFA) uses data, models, and analytics to assess public health threats. We use a range of quantitative and qualitative techniques in our assessments, which help support public health response.
One of our approaches is to develop outbreak scenario assessments that consider a range of possibilities in the coming weeks to months of an outbreak. To develop scenarios, we consider insights from previous outbreaks of the disease, the current state of the outbreak, and the potential role of public health interventions in the outbreak. We then identify indicators that help to determine which of the possible scenarios is most likely to occur. We also incorporate modeling insights to highlight key sources of uncertainty or address questions from public health practitioners or decision-makers. CFA's scenario assessments inform CDC leadership, public health professionals, and partners engaged in planning and executing emergency response activities.
Our Methods
To conduct an outbreak scenario assessment, we assemble a team of experts, including subject-matter experts from disease-specific programs, as well as those specializing in surveillance and outbreak dynamics. Outbreak scenarios are intended to be future-focused; they outline plausible scenarios for what could occur in the coming weeks to months. To identify potential outbreak scenarios, we consider a range of inputs including past and current outbreak characteristics, such as outbreak duration, geographic spread, populations affected, age structure, and transmission settings.
To assess which scenario is most likely occurring, we provide indicators and associated criteria for each indicator that correspond to each scenario. Indicators are not exhaustive, and not all indicators need to be met for the assessment of a given scenario. We rely on expert judgment to determine which scenario is most likely occurring. The list of indicators for an outbreak scenario assessment may include:
- Detection of cases in new populations or geographic regions
- Measures of outbreak growth or containment
- Extent and/or success of interventions, such as the proportion of contacts identified through contact tracing or the proportion of contacts in quarantine
- Disease-specific characteristics, such as clinical severity
After examining available evidence and expert judgment for each indicator, we identify which level (i.e., scenario) each indicator currently falls under. We display a table with the list of indicators and currently assessed level for each indicator (Figure 1). When evidence is uncertain, we can assign multiple levels to an indicator. After assigning a level for each indicator, we rely on expert judgment to determine the most likely future scenario. In some cases, the most likely future scenario is the scenario supported by evidence for the majority of indicators; in others, there is one driving indicator that is deemed more important. We also assign confidence levels to each of our assessments, based on the quality and amount of evidence that underpins the assessment, as well as how well different lines of evidence corroborate one another.
Identifying and incorporating modeling insights
Insights from modeling can be incorporated into outbreak scenario assessments to address questions decision-makers may be interested in for an outbreak response, such as:
- Which intervention would be most effective?
- How large will the outbreak be? How far into the outbreak are we?
- What are key sources of uncertainty in outbreak? How much will known uncertainties matter in progression of outbreak?
Depending on the availability of data and modeling tools for a given question, scenario assessments can begin as qualitative products and expand to include quantitative results as they become available. When data are limited, a scenario assessment summarizes insights from literature and applies model-informed thinking to help answer one or more questions. Insights from simple models can also be applied to illustrate the role of key uncertainties or policy decisions. As more data becomes available and models are stood up using outbreak data, the scenario assessment may integrate quantitative data and modeling results that are more tailored to the outbreak at hand.