Coronavirus Forecast

Why Estimated COVID-19 Death in the US is Critical

The CDC responds to the 2019 coronavirus pandemic (COVID-19) caused by a new coronavirus, SARS-CoV-2, which spreads from person to person. The federal government works closely with state, tribal, local and territorial health departments, and other public health partners, to respond to this situation. Death estimates will help inform public health decision-making by projecting impacts that might occur in the coming weeks.

What Predictions Predict forecasts

Forecast based on the use of statistical or mathematical models (hereinafter referred to as "models") aims to predict cumulative changes at the national and state levels that report COVID-19 deaths over the next four weeks. The forecasting team estimates the number of deaths using various types of data (e.g., COVID-19 data, demographic data, mobility data), methods (see below), and estimated impact of interventions (eg social distance, use of face covering).

National Forecast

• This forecast shows the cumulative reporting of COVID-19 deaths since February and estimated deaths over the next four weeks in the United States.
• The CU model makes various assumptions about the effectiveness of current interventions. All other models assume that existing social distance measures will continue for the indicated time period.

Forecast Interpretation

• National-level forecasts indicate that mortality is likely to continue to increase in the coming weeks. How fast they will increase is still very uncertain.
• Models that include strong contact reductions (eg, IHME, MOBS) suggest that new deaths will continue to occur, but slow down substantially over the next four weeks. In contrast, models that do not include strong contact reductions (eg CU 20%, CU 30%) suggest that total deaths can continue to increase rapidly.
• Country-level estimates vary widely, reflecting differences in the initial epidemic phase, timing of interventions, and model-specific assumptions (eg regarding the effectiveness of interventions applied).

Country Forecast

Country-level estimates show COVID-19 cumulative deaths observed and estimated state levels in the US.

Forecasts fall into one of three categories

• The LANL model does not explicitly model the effects of individual distance social measures but assumes that implemented interventions will continue to be enforced in the future which results in decreased growth.
• The IHME, UT Austin, University of Geneva, YYG, and MOBS_NEU models depend on existing social distance measures that continue throughout the projection period.
• The CU model makes different assumptions about the effectiveness of current social distance interventions.

Download the estimated PDF

Download the data excel

Work to Unite Forecasts for COVID-19 Death in the US

CDC works with partners to put together weekly forecasts for COVID-19 deaths in one place. This forecast has been independently developed and shared publicly. It is important to bring together these forecasts to help understand how they compare with each other and how much uncertainty about what might happen in the next four weeks.

Columbia University

Model name: 20% CU contact reduction, 30% CU contact reduction, 40% CU contact reduction

Assumption of intervention
These models are based on the assumption of reducing the number of contacts per case. Three different contact reduction adaptation scenarios are projected: contact reduction of 20%, 30%, and 40% in US countries with at least 10 cases. Additional reductions are implemented with the addition of new cases, and all social distance interventions remain until the end of the projection.

Method
SEAP metapopulation model