Below, we explain the methods behind the Center for Forecasting and Outbreak Analytics' rapid risk assessments. Subject-matter experts make qualitative judgments based on available evidence in an evolving situation. For each population examined, we estimate risk by combining the likelihood and impact of infection at a population level.
Overview
ÐÇ¿ÕÓéÀÖ¹ÙÍø'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 and strengthen national security.
One of our qualitative approaches is a population-based rapid risk assessment where subject-matter experts make judgements based on available evidence, including epidemiological and clinical data, modeling analyses, and data on past outbreaks.
We continue to develop methods for other qualitative approaches to assessing risk, including scenario-based assessments, either as alternatives to or in combination with population-based rapid risk assessments.
Our Methods
To conduct a rapid risk assessment, we assemble a team of experts, including risk assessors and subject-matter experts from disease-specific programs, as well as those specializing in surveillance and outbreak dynamics. Before starting a risk assessment, we consider framing questions such as:
Who: What populations are currently affected or at risk?
What: What is the disease or pathogen of concern?
When: What is the timeframe for this risk assessment?
Where: Where is the event occurring right now?
After conducting a review of the literature and available evidence, the risk assessment team identifies the population(s) to include in the assessment. We consider the risk to the general (overall) population in all risk assessments and separately assess specific populations at higher risk or of importance to public health practitioners or decision-makers. As an outbreak evolves, there are often particular populations of focus, given historical context or current information about the outbreak. We prioritize populations with a higher likelihood of infection, lower population immunity to infection or severe disease, or higher degrees of impact. We note that while we may prioritize specific populations, they also are part of the general population.
For each population examined, we estimate risk by combining the likelihood and impact of infection at a population level (Figure 1). The risk is higher for a given population as the likelihood and/or impact increases.
Figure 1: Risk assessors combine the likelihood of infection and the impact of the disease to estimate the risk posed to a given population.
Matrix chart of risk levels when comparing likelihood of becoming infected with the impact of the disease on the population. When likelihood and impact are extremely low so is the risk. When likelihood and/or impact are high or very high, the risk can range from moderate to very high levels. The risk assessment tool illustrates a direct or positive correlation.
The full matrix set compares impact to likelihood of infection to determine overall risk.
When impact is extremely low.
Extremely low likelihood is extremely low risk.
Very low likelihood is very low risk.
Low likelihood is very low risk.
Moderate likelihood is low risk.
High likelihood is low risk.
Very high likelihood is moderate risk.
When impact is very low
Extremely low likelihood is very low risk.
Very low likelihood is very low risk.
Low likelihood is low risk.
Moderate likelihood is low risk.
High likelihood is moderate risk.
Very high likelihood is moderate risk.
When impact is low
Extremely low likelihood is very low risk.
Very low likelihood is low risk.
Low likelihood is low risk.
Moderate likelihood is moderate risk.
High likelihood is moderate risk.
Very high likelihood is high risk.
When impact is moderate
Extremely low likelihood is low risk.
Very low likelihood is low risk.
Low likelihood is moderate risk.
Moderate likelihood is moderate risk.
High likelihood is high risk.
Very high likelihood is high risk.
When impact is high
Extremely low likelihood is low risk.
Very low likelihood is moderate risk.
Low likelihood is moderate risk.
Moderate likelihood is high risk.
High likelihood is high risk.
Very high likelihood is very high risk.
When impact is very high
Extremely low likelihood is moderate risk.
Very low likelihood is moderate risk.
Low likelihood is high risk.
Moderate likelihood is high risk.
High likelihood is very high risk.
Very high likelihood is very high risk.
The confidence level given for this assessment can be low, moderate, or high.
We assess the likelihood and impact using a set of definitions (see below); however, definitions are not all-inclusive, and the risk assessment team assigns likelihood and impact levels based on a subjective assessment of available information. In some cases, the likelihood level is a combination of all factors assessed; in others, there is one driving factor that is deemed more important.
The likelihood of infection refers to the probability or extent of infection across the population of interest, which depends on the likelihood of infection in the population, number of people infected, level of population immunity against infection and current response capacity to limit spread, such as contact tracing or availability of testing (see definition tables).The impact of infection considers factors such as the severity of disease, level of population immunity against severe outcomes, and current resources available to limit the impact of infection, such as medical countermeasures including vaccination or treatments (see definition tables). 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 (see definition tables).
After assessing the risk for each population, we summarize key factors that could change our assessment. For some public health threats or pathogens, we recognize that certain factors could rapidly and substantially change the risk posed to populations assessed in a rapid risk assessment. Therefore, future risk is assessed in a subsequent section, if applicable.
Our risk assessment methods are adapted from those developed by the and the .
Likelihood of infection depends on factors such as the number of people infected, pathogen infectiousness and population immunity against infection, population prevalence of infection, and the response capacity to limit spread.
Likelihood of infection.
Probability or extent of infection across population of interest.
Note: Infection likelihood may cross multiple probability levels between assessment factors, and some factors may be prioritized over others.
Extremely low: An extremely small number of people are likely to be exposed and/or infected, the pathogen is not very infectious, or the population is highly immune to infection. We expect an extremely low prevalence of infection in the population, far less than 1% of the population. There are very high levels of resources available to limit spread.
Very low: A very small number of people are likely to be exposed and/or infected, the pathogen is not very infectious, or the population is highly immune to infection. We expect a very low prevalence of infection in the population. There are high levels of resources available to limit spread.
Low: There are limited opportunities for exposure for most of the population, but exposure and/or infection may be high in some areas or subgroups. The pathogen has at least moderate infectiousness or significant gaps in population immunity to infection. We expect a low prevalence of infection in the population, potentially with pockets of higher prevalence. There are low levels of resources available to limit spread, but availability of resources may be higher in some areas or subgroups.
Moderate: Many people are likely to be exposed and/or infected, the pathogen has moderate to high infectiousness, or the population has low levels of immunity to infection. We expect a moderate prevalence of infection in the population. There are moderate levels of resources available to limit spread, and availability of resources may differ across areas or subgroups.
High: Most people are likely to be exposed and/or infected, the pathogen has high infectiousness, or the population has very low immunity to infection. We expect a high prevalence of infection in the population, with most of the population affected. There are very few resources available to limit spread.
Very high: The vast majority of the population is likely to be exposed, the pathogen has very high infectiousness, or the population has extremely low immunity to infection. We expect a very high prevalence of infection in the population, with the vast majority of the population affected. There are extremely limited resources available to limit spread.
Impact of infection depends on factors such as disease severity, population immunity against severe disease, resources available to limit spread, and societal impact.
Impact of infection.
Collective impact of individual infections across the population, as well as population-wide impacts.
Note: Infection impact may cross multiple severity levels between assessment factors, and some factors may be prioritized over others.
Extremely low: The pathogen is extremely unlikely to cause severe disease for this population or there is an extremely high proportion of population immunity protecting against severe disease. There are near to unlimited levels of resources available to limit impact. The disease is expected to cause little or no disruption to normal activities and require little to no additional resources for public health measures.
Very low: The pathogen is very unlikely to cause severe disease for this population or there is a very high proportion of population immunity protecting against severe disease. There are very high levels of resources available to limit impact. The disease is very unlikely to cause disruption to normal activities or require additional resources for public health measures.
Low: The pathogen is unlikely to cause severe disease for this population or there is a high proportion of population immunity protecting against severe disease. There are high levels of resources available to limit impact. The disease is unlikely to cause disruption to normal activities or require additional resources for public health measures.
Moderate: The pathogen causes severe disease for a substantial proportion of this population or pockets within this population or there is limited population immunity protecting people from severe disease. There are moderate levels of resources available to limit impact, and availability of resources may differ across areas or subgroups. The disease may cause significant disruption to the population and require significant public health resources.
High: The pathogen typically causes severe disease for this population or there is a very low proportion of population immunity protecting against severe disease. There are very few resources available to limit impact. The disease could cause extensive disruption to normal activities and will potentially require a large amount of public health resources.
Very high: The pathogen typically causes very severe disease for this population or there is a very low proportion of population immunity protecting against severe disease. There are extremely limited resources available to limit impact. The disease could cause prolonged and extensive disruption to normal activities and will potentially require a very high level of public health resources.
The confidence level of an assessment depends on information availability and any information gaps or assumptions.
Confidence level of assessment.
Confidence of assessment based on the extent and quality of information and how well different lines of evidence corroborate one another.
Low confidence: Assessment is based on information that is fragmented, poorly corroborated, or upon data sources for which there are significant concerns or problems. There may be several information gaps that require numerous assumptions in order to draw conclusions for the assessment.
Moderate confidence: Assessment is based on credibly sourced and plausible information, but the information is not of sufficient quality or corroboration to warrant a high level of confidence. The assessment acknowledges some information gaps or assumptions that underlie analysis.
High confidence: Assessment is based on high-quality information from multiple sources, although such judgments are not a certainty. There are few information gaps, and few assumptions are required to draw analytic conclusions.
