Model Projection

New scenario for models are defined in each round

Round 1

Scenario defined as of 2020-12-22
Model Projecting from Epiweek 53 to Epiweek 26



Special regions (American Samoa, Guam, Northern Marianas Island, Virgin Islands) not included
Loading...

Definitions


NPI: NonPharmaceutical Interventions (e.g. masks, social distancing)

Epiweek: Epidemiological Week as defined by MMWR

Notes:

Ensemble methods: The Scenario Modeling Hub ensembles individual projections using one method: Ensemble is obtained by calculating the median of each submitted quantile.
Ensemble projection includes only those submissions that reported all quantiles and the four scenarios for their targets. Individual model and ensemble projections are available in the GitHub Repository.

Ground truth data: The model projections for cumulative hospitalizations start at zero, so no observed data are shown.

Licensing: Models projection are available by default under a CC BY 4.0 license. Some models have specific license. See repository for details.

Disclaimer: The content of the COVID-19 Scenario Modeling Hub is solely the responsability of the participating teams and the Hub maintainers and does not represent the official views of any related funding organizations


Model Projection

New scenario for models are defined in each round

Round 2

Scenario defined as of 2021-01-22
Model Projecting from Epiweek 3 to Epiweek 29



Special regions (American Samoa, Guam, Northern Marianas Island, Virgin Islands) not included
Loading...

Definitions


NPI: NonPharmaceutical Interventions (e.g. masks, social distancing)

Epiweek: Epidemiological Week as defined by MMWR

Notes:

Ensemble methods: The Scenario Modeling Hub ensembles individual projections using one method: Ensemble is obtained by calculating the median of each submitted quantile.
Ensemble projection includes only those submissions that reported all quantiles and the four scenarios for their targets. Individual model and ensemble projections are available in the GitHub Repository.

Ground truth data: The model projections for cumulative hospitalizations start at zero, so no observed data are shown.

Licensing: Models projection are available by default under a CC BY 4.0 license. Some models have specific license. See repository for details.

Disclaimer: The content of the COVID-19 Scenario Modeling Hub is solely the responsability of the participating teams and the Hub maintainers and does not represent the official views of any related funding organizations


Model Projection

New scenario for models are defined in each round

Round 3

Scenario defined as of 2021-03-05
Model Projecting from Epiweek 9 to Epiweek 35



Special regions (American Samoa, Guam, Northern Marianas Island, Virgin Islands) not included
Loading...

Definitions


NPI: NonPharmaceutical Interventions (e.g. masks, social distancing)

Epiweek: Epidemiological Week as defined by MMWR

Notes:

Ensemble methods: The Scenario Modeling Hub ensembles individual projections using one method: Ensemble is obtained by calculating the median of each submitted quantile.
Ensemble projection includes only those submissions that reported all quantiles and the four scenarios for their targets. Individual model and ensemble projections are available in the GitHub Repository.

Ground truth data: The model projections for cumulative hospitalizations start at zero, so no observed data are shown.

Projection truncation: Since models were not required to incorporate behavioral feedbacks that would likely accompany large increases in COVID-19 cases, hospitalizations, or deaths, we truncate projected values beyond 80% of previously seen peaks. Within each location, truncation is evaluated separately for incident cases and deaths.
The truncation threshold is calculated as 80% of the highest observed weekly incidence, adjusted by population size. Once either threshold is crossed, all projections (incident and cumulative case, death, and hospitalization) are truncated using following the following rules:

  - If an individual model projects incident cases or deaths to be higher than the threshold, median projections are not plotted for that model beyond the week at which the threshold is crossed. Uncertainty intervals are retained for all weeks. This truncation is applied to all states, except those for which truncation would remove all projection points.
  - If the Ensemble crosses the threshold, no projections (median or uncertainty intervals) are plotted beyond the first week at which the threshold is crossed.

Licensing: Models projection are available by default under a CC BY 4.0 license. Some models have specific license. See repository for details.

Disclaimer: The content of the COVID-19 Scenario Modeling Hub is solely the responsability of the participating teams and the Hub maintainers and does not represent the official views of any related funding organizations


Model Projection

New scenario for models are defined in each round

Round 4

Scenario defined as of 2021-03-28
Model Projecting from Epiweek 12 to Epiweek 38



Special regions (American Samoa, Guam, Northern Marianas Island, Virgin Islands) not included
Loading...

Definitions


NPI: NonPharmaceutical Interventions (e.g. masks, social distancing)

Epiweek: Epidemiological Week as defined by MMWR

Notes:

Ensemble methods: The Scenario Modeling Hub ensembles individual projections using one method: Ensemble is obtained by calculating the median of each submitted quantile.
Ensemble projection includes only those submissions that reported all quantiles and the four scenarios for their targets. Individual model and ensemble projections are available in the GitHub Repository.

Ground truth data: The model projections for cumulative hospitalizations start at zero, so no observed data are shown.

Projection truncation: Since models were not required to incorporate behavioral feedbacks that would likely accompany large increases in COVID-19 cases, hospitalizations, or deaths, we truncate projected values beyond 80% of previously seen peaks. Within each location, truncation is evaluated separately for incident cases and deaths.
The truncation threshold is calculated as 80% of the highest observed weekly incidence, adjusted by population size. Once either threshold is crossed, all projections (incident and cumulative case, death, and hospitalization) are truncated using following the following rules:

  - If an individual model projects incident cases or deaths to be higher than the threshold, median projections are not plotted for that model beyond the week at which the threshold is crossed. Uncertainty intervals are retained for all weeks. This truncation is applied to all states, except those for which truncation would remove all projection points.
  - If the Ensemble crosses the threshold, no projections (median or uncertainty intervals) are plotted beyond the first week at which the threshold is crossed.

Licensing: Models projection are available by default under a CC BY 4.0 license. Some models have specific license. See repository for details.

Disclaimer: The content of the COVID-19 Scenario Modeling Hub is solely the responsability of the participating teams and the Hub maintainers and does not represent the official views of any related funding organizations


Model Projection

New scenario for models are defined in each round

Round 5

Scenario defined as of 2021-05-02
Model Projecting from Epiweek 17 to Epiweek 43



Special regions (American Samoa, Guam, Northern Marianas Island, Virgin Islands) not included
Loading...

Definitions


NPI: NonPharmaceutical Interventions (e.g. masks, social distancing)

Epiweek: Epidemiological Week as defined by MMWR

LOP: Linear Opinion Pool; method used to calculate Ensemble_LOP by averaging cumulative probabilities of a given value across submissions
Moderate NPI
(moderate NPI reduction)
Low NPI
(high NPI reduction)
High Vaccination
(low hesistancy)
Scenario A
Vaccination
Coverage saturates at 83% nationally among the vaccine-eligible population* by October 31, 2021**


NPIs
50% reduction from April 2021 to October 2021

Scenario B
Vaccination
Coverage saturates at 83% nationally among the vaccine-eligible population* by October 31, 2021**


NPIs
80% reduction from April 2021 to October 2021

Low Vaccination
(high hesistancy)
Scenario C
Vaccination
Coverage saturates at 68% nationally among the vaccine-eligible population* by October 31, 2021**


NPIs
50% reduction from April 2021 to October 2021

Scenario D
Vaccination
Coverage saturates at 68% nationally among the vaccine-eligible population* by October 31, 2021**


NPIs
80% reduction from April 2021 to October 2021



* Vaccine-eligible population. The eligible population for vaccination is presumed to be individuals aged 16 years or older until June 1, 2021. On June 1, the eligible population is presumed to extend to individuals aged 12 years and older.

** Vaccine hesitancy is expected to cause vaccination coverage to slow and eventually saturate at some level below 100%. The saturation levels provided in these scenarios are National reference points to guide defining hesitancy, though the speed of that saturation and heterogeneity between states (or other geospatial scales) and/or age groups are at the discretion of the modeling team. The high vaccination 83% saturation is defined using the current estimates from the Delphi group (link) from March 13, 2021 data. The low saturation estimate of 68% is the lowest county-level estimate from the U.S. Census Bureau’s Pulse Survey from March 15, 2021 data (link). Both of these saturation levels are assumed to be among the population eligible for vaccination, not the full population.

Notes:

Ensemble methods: Currently, the Scenario Modeling Hub ensembles individual projections using two methods:
(a) Ensemble_LOP is calculated by averaging cumulative probabilities of a given value across submissions.
(b) Ensemble is obtained by calculating the median of each submitted quantile.
Ensembles projection include only those submissions that reported all quantiles and the four scenarios for their targets. Individual model and ensemble projections are available in the GitHub Repository.

Ground truth data: The model projections for cumulative hospitalizations start at zero, so no observed data are shown.

Projection truncation: Since models were not required to incorporate behavioral feedbacks that would likely accompany large increases in COVID-19 cases, hospitalizations, or deaths, we truncate projected values beyond 80% of previously seen peaks. Within each location, truncation is evaluated separately for incident cases and deaths.
The truncation threshold is calculated as 80% of the highest observed weekly incidence, adjusted by population size. Once either threshold is crossed, all projections (incident and cumulative case, death, and hospitalization) are truncated using following the following rules:

  - If an individual model projects incident cases or deaths to be higher than the threshold, median projections are not plotted for that model beyond the week at which the threshold is crossed. Uncertainty intervals are retained for all weeks. This truncation is applied to all states, except those for which truncation would remove all projection points.
  - If the LOP Ensemble crosses the threshold, no projections (median or uncertainty intervals) are plotted beyond the first week at which the threshold is crossed.

Licensing: Models projection are available by default under a CC BY 4.0 license. Some models have specific license. See repository for details.

Disclaimer: The content of the COVID-19 Scenario Modeling Hub is solely the responsability of the participating teams and the Hub maintainers and does not represent the official views of any related funding organizations


Model Projection

New scenario for models are defined in each round

Round 6

Scenario defined as of 2021-05-25
Model Projecting from Epiweek 21 to Epiweek 47



Special regions (American Samoa, Guam, Northern Marianas Island, Virgin Islands) not included
Loading...

Definitions


NPI: NonPharmaceutical Interventions (e.g. masks, social distancing)

Epiweek: Epidemiological Week as defined by MMWR

LOP: Linear Opinion Pool; method used to calculate Ensemble_LOP by averaging cumulative probabilities of a given value across submissions
Low Impact Variant
(low transmissibility increase, no immune escape)
High Impact Variant
(high transmissibility increase, no immune escape)
High Vaccination
(low hesistancy)
Scenario A
Vaccination
- Coverage saturates at 86% nationally among the vaccine-eligible population* by November 30, 2021**
- VE is 50%/90% for Pfizer/Moderna against currently circulating variants (1st/2nd dose)
- J&J no longer used


Variant
20% increased transmissibility as compared by B.1.1.7 for B.1.617+ variant. 5% prevalence of B.1.617+ nationally on May 29.

Scenario B
Vaccination
- Coverage saturates at 86% nationally among the vaccine-eligible population* by November 30, 2021**
- VE is 50%/90% for Pfizer/Moderna against currently circulating variants (1st/2nd dose)
- J&J no longer used


Variant
60% increased transmissibility as compared by B.1.1.7 for B.1.617+ variant. 5% prevalence of B.1.617+ nationally on May 29.

Low Vaccination
(high hesistancy)
Scenario C
Vaccination
- Coverage saturates at 75% nationally among the vaccine-eligible population* by November 30, 2021**
- VE is 50%/90% for Pfizer/Moderna against currently circulating variants (1st/2nd dose) and 60% for JJ (1 dose)
- J&J no longer used


Variant
20% increased transmissibility as compared by B.1.1.7 for B.1.617+ variant. 5% prevalence of B.1.617+ nationally on May 29.

Scenario D
Vaccination
- Coverage saturates at 75% nationally among the vaccine-eligible population* by November 30, 2021**
- VE is 50%/90% for Pfizer/Moderna against currently circulating variants (1st/2nd dose) and 60% for JJ (1 dose)
- J&J no longer used


Variant
60% increased transmissibility as compared by B.1.1.7 for B.1.617+ variant. 5% prevalence of B.1.617+ nationally on May 29.



* Vaccine-eligible population. The eligible population for vaccination is presumed to be individuals aged 16 years or older until May 12, 2021. On May 12, the eligible population is extended to individuals aged 12 years and older, through the end of the projection period.

** Vaccine hesitancy expected to cause vaccination coverage to slow and eventually saturate at some level below 100%. The saturation levels provided in these scenarios are National reference points to guide defining hesitancy, though the speed of that saturation and heterogeneity between states (or other geospatial scales) and/or age groups are at the discretion of the modeling team. The high vaccination 86% saturation is defined using the current estimates from the Delphi group (link, updated from Round 5). The low saturation estimate of 75% is the lowest county-level estimate from the U.S. Census Bureau’s Pulse Survey from Apr 14-26, 2021 data (link), which is updated from Round 5

Notes:

Ensemble methods: Currently, the Scenario Modeling Hub ensembles individual projections using two methods:
(a) Ensemble_LOP is calculated by averaging cumulative probabilities of a given value across submissions.
(b) Ensemble is obtained by calculating the median of each submitted quantile.
Ensembles projection include only those submissions that reported all quantiles and the four scenarios for their targets. Individual model and ensemble projections are available in the GitHub Repository.

Ground truth data: The model projections for cumulative hospitalizations start at zero, so no observed data are shown.

Projection truncation: Since models were not required to incorporate behavioral feedbacks that would likely accompany large increases in COVID-19 cases, hospitalizations, or deaths, we truncate projected values beyond 80% of previously seen peaks. Within each location, truncation is evaluated separately for incident cases and deaths.
The truncation threshold is calculated as 80% of the highest observed weekly incidence, adjusted by population size. Once either threshold is crossed, all projections (incident and cumulative case, death, and hospitalization) are truncated using following the following rules:

  - If an individual model projects incident cases or deaths to be higher than the threshold, median projections are not plotted for that model beyond the week at which the threshold is crossed. Uncertainty intervals are retained for all weeks. This truncation is applied to all states, except those for which truncation would remove all projection points.
  - If the LOP Ensemble crosses the threshold, no projections (median or uncertainty intervals) are plotted beyond the first week at which the threshold is crossed.

Licensing: Models projection are available by default under a CC BY 4.0 license. Some models have specific license. See repository for details.

Disclaimer: The content of the COVID-19 Scenario Modeling Hub is solely the responsability of the participating teams and the Hub maintainers and does not represent the official views of any related funding organizations


Model Projection

New scenario for models are defined in each round

Round 7

Scenario defined as of 2021-07-06
Model Projecting from Epiweek 27 to Epiweek 1



Special regions (American Samoa, Guam, Northern Marianas Island, Virgin Islands) not included
Loading...

Definitions


NPI: NonPharmaceutical Interventions (e.g. masks, social distancing)

Epiweek: Epidemiological Week as defined by MMWR

LOP: Linear Opinion Pool; method used to calculate Ensemble_LOP by averaging cumulative probabilities of a given value across submissions
Low Impact Variant
(low transmissibility increase, no immune escape)
High Impact Variant
(high transmissibility increase, no immune escape)
High Vaccination
(low hesistancy)
Scenario A
Vaccination
- Coverage saturates at 80% nationally among the vaccine-eligible population* by December 31, 2021**
- VE is 50%/90% for Pfizer/Moderna against the Delta variant, against symptoms (1st/2nd dose)
- J&J no longer used


Variant
40% increased transmissibility as compared with Alpha for Delta variant. Initial prevalence estimated at state-level by teams.

Scenario B
Vaccination
- Coverage saturates at 80% nationally among the vaccine-eligible population* by December 31, 2021**
- VE is 35%/85% for Pfizer/Moderna against the Delta variant, against symptoms (1st/2nd dose)
- J&J no longer used


Variant
60% increased transmissibility as compared with Alpha for Delta variant. Initial prevalence estimated at state-level by teams.

Low Vaccination
(high hesistancy)
Scenario C
Vaccination
- Coverage saturates at 70% nationally among the vaccine-eligible population* by December 31, 2021**
- VE is 50%/90% for Pfizer/Moderna against the Delta variant, against symptoms (1st/2nd dose)
- J&J no longer used


Variant
40% increased transmissibility as compared with Alpha for Delta variant. Initial prevalence estimated at state-level by teams.

Scenario D
Vaccination
- Coverage saturates at 70% nationally among the vaccine-eligible population* by December 31, 2021**
- VE is 35%/85% for Pfizer/Moderna against the Delta variant, against symptoms (1st/2nd dose)
- J&J no longer used


Variant
60% increased transmissibility as compared with Alpha for Delta variant. Initial prevalence estimated at state-level by teams.



* Vaccine-eligible population. The eligible population for vaccination is presumed to be individuals aged 12 years and older through the end of the projection period.

** Vaccine hesitancy is expected to cause vaccination coverage to slow and eventually saturate at some level below 100%. The saturation levels provided in these scenarios are National reference points to guide defining hesitancy, though the speed of that saturation and heterogeneity between states (or other geospatial scales) and/or age groups are at the discretion of the modeling team. The high vaccination 80% saturation is defined crudely as using the current estimates from the Delphi group, adjusted for potential bias in respondents, who tend to be more highly vaccinated that the general US population (link, updated from Round 6). The low saturation estimate of 70% is based on an adjustment of the Pulse Survey overall estimate, adjusted for survey participant vaccination coverage. This number also mirrors the lowest county-level estimate (73.3%) from the U.S. Census Bureau’s Pulse Survey from May 26-June 7, 2021 data (link), which is updated from Round 6.

Notes:

Ensemble methods: Currently, the Scenario Modeling Hub ensembles individual projections using two methods:
(a) Ensemble_LOP is calculated by averaging cumulative probabilities of a given value across submissions.
(b) Ensemble is obtained by calculating the median of each submitted quantile.
Ensembles projection include only those submissions that reported all quantiles and the four scenarios for their targets. Individual model and ensemble projections are available in the GitHub Repository.

Ground truth data: The model projections for cumulative hospitalizations start at zero, so no observed data are shown.

Projection truncation: Since models were not required to incorporate behavioral feedbacks that would likely accompany large increases in COVID-19 cases, hospitalizations, or deaths, we truncate projected values beyond 80% of previously seen peaks. Within each location, truncation is evaluated separately for incident cases and deaths.
The truncation threshold is calculated as 80% of the highest observed weekly incidence, adjusted by population size. Once either threshold is crossed, all projections (incident and cumulative case, death, and hospitalization) are truncated using following the following rules:

  - If an individual model projects incident cases or deaths to be higher than the threshold, median projections are not plotted for that model beyond the week at which the threshold is crossed. Uncertainty intervals are retained for all weeks. This truncation is applied to all states, except those for which truncation would remove all projection points.
  - If the LOP Ensemble crosses the threshold, no projections (median or uncertainty intervals) are plotted beyond the first week at which the threshold is crossed.

Licensing: Models projection are available by default under a CC BY 4.0 license. Some models have specific license. See repository for details.

Disclaimer: The content of the COVID-19 Scenario Modeling Hub is solely the responsability of the participating teams and the Hub maintainers and does not represent the official views of any related funding organizations


Loading...

Scenario Definition


Loading...


Loading...


Scenario Definition


Loading...


Loading...


Scenario Definition


Loading...


Loading...


Scenario Definition


Loading...


Loading...

Common Specifications

Vaccination

  • Doses available:
    • 50M Moderna/Pfizer 1st doses, 15M J&J doses available monthly
    • Supply has likely eclipsed demand at this stage. Number of doses are for reference and as a reminder to account for different VE by manufacturer, but no longer indicate number of doses administered. Distribution of doses by manufacturer and associated vaccine efficacy should fit within these dose bounds.
  • VE:
    • 75% and 95% against symptoms (Moderna/Pfizer 1st and 2nd dose)
    • 70% against symptoms (J&J)
    • VE is defined here as vaccine effectiveness against symptomatic disease. Teams should make their own informed assumptions about effectiveness and impacts on other outcomes (e.g., infection, hospitalization, death). Data on VE studies of infection and symptomatic disease are included below.

Vaccination Hesitancy

Vaccine hesitancy is expected to cause vaccination coverage to slow and eventually saturate at some level below 100%. The saturation levels provided in these scenarios are illustrative National reference points to guide defining hesitancy. The high vaccination scenario (low hesitancy) saturates at approximately 83% vaccination coverage nationally among the eligible population, as defined by current estimates from the Delphi group (link) from March 13, 2021 data (red line in figure). The low vaccination scenario (high hesitancy) saturates at approximately 68% vaccination coverage nationally among the eligible population, defined by the lowest county-level estimate from the U.S. Census Bureau’s Pulse Survey (link) from March 15, 2021 data. The speed of vaccination saturation should be defined by the modeling team, and can be defined as a logistic function (red and blue lines in figure below) or at different speeds (green line below). State or smaller geospatial unit and/or age group hesitancy limits should be defined by the modeling team using their best judgement. Overall national hesitancy should be similar to the illustrative levels defined in the scenarios, though is not expected to be exact. The eligible population for vaccination is presumed to be individuals aged 16 years or older until June 1, 2021. On June 1, the eligible population is presumed to extend to individuals aged 12 years and older.

Submission Information

Scenario Scenario name for submission file Scenario ID for submission file
Scenario A. High Vaccination, Moderate NPI highVac_modNPI A-2021-05-02
Scenario B. High Vaccination, Low NPI highVac_lowNPI B-2021-05-02
Scenario C. Low Vaccination, Moderate NPI lowVac_modNPI C-2021-05-02
Scenario D. Low Vaccination & Low NPI lowVac_lowNPI D-2021-05-02
  • Submission date: May 4, 2021
  • End date for fitting data: May 1, 2021 - no fitting should be done to data from after this date
  • Start date for fifth-round scenarios: May 2, 2021 (week ending May 8) - first date of simulated outcomes
  • Simulation end date: October 30, 2021 (26-week horizon)

Scenario and Simulation Details

  • Social Distancing Measures:
    • Includes combined effectiveness/impact of all non-pharmaceutical interventions and behavior change.
    • Declines over a period of 6 months starting at the start of May 2021 and ending in October 2021 at designated level of control effectiveness observed for April 2021.
    • Decline can be implemented at teams’ discretion (e.g., daily or monthly stepdowns).
    • Decline can be gradual or sudden, and can differ in speed between states.
    • The effectiveness of control in April 2021 should be based on the last two weeks of the month.
    • Reduction should be implemented based on each team’s best judgment, but should be done in such a way that a 100% reduction (0% of April 2021 effectiveness) would approximate an epidemic without NPIs (e.g., no masks, no social distancing) in place, but still including immunity, vaccination, etc. We recognize there is uncertainty about what transmission would be without NPIs; this uncertainty should be incorporated into the scenario projections.
  • Testing-Trace-Isolate: constant at baseline levels
  • Masking: Included as part of “Social Distancing Measures” above.
  • Vaccination:
    • Pfizer / Moderna
      • Vaccine efficacy (2-dose vaccines):
        • First dose: 75% against symptoms, 14 days after 1st dose
        • Second dose: 95% against symptoms, 14 days after 2nd dose
        • Effectiveness and impact on infection and other outcomes (hospitalizations, deaths) is at team’s discretion and should be clearly documented in team’s metadata.
        • Doses 3.5 weeks apart
      • Vaccine availability:
        • December-April: based on data on administered doses
        • May-October: 50 million available first doses/month, with the intention of protocols being followed (100M doses/mo)
    • Johnson & Johnson
      • Vaccine efficacy (1-dose vaccine):
        • Single dose: 70% against symptoms, 14 days after dose
        • Effectiveness and impact on infection and other outcomes (hospitalizations, deaths) is at team’s discretion and should be clearly documented in team’s metadata.
      • Vaccine availability:
        • March-April: based on data on administered doses, with continuing at rate current on date of projection for remainder of month.
        • April-September: 15M doses/month available May-October 2021
  • Vaccine Hesitancy: Vaccine hesitancy expected to cause vaccination coverage to slow and saturate below 100%. National vaccination saturation levels designated for each scenario serve as illustrative reference points to guide defining hesitancy, though the speed of that saturation and heterogeneity between states (or other geospatial scale) and/or age groups are at the discretion of the team.
  • B.1.1.7 Variant strain: Teams should model the B.1.1.7 variant as appropriate to their model. Any assumptions (e.g., differences in severity/mortality, VE, or natural immunity) should be clearly defined in the metadata. The default assumptions are that the variant is 1.5x more transmissible than current strains and reaches 50% dominance by March 15 and 100% dominance by May 1 (see MMWR report); here a 1.5x increase in transmissibility is defined as the increase in the expected number of infections by a single infected individual over their entire course of infection when there are no interventions or immunity in the population (e.g., a 1.5x increase in R0 in a classic epidemic model). No differences between B.1.1.7 and current strains in severity, mortality, or VE are assumed in default.
  • Transmission assumptions: models fit to US state-specific dynamic up until End date for fitting data specified above – no proscribed R0, interventions, etc.
  • Pathogenicity assumptions: no exogenous fluctuations in pathogenicity/transmissibility beyond seasonality effects
  • Vaccine effectiveness: level of effectiveness and available doses are specified for each scenario; assumptions regarding time required to develop immunity, age-related variation in effectiveness, duration of immunity, and additional effects of the vaccine on transmission are left to the discretion of each team
  • Vaccine allocation: between-state allocation is based on population per the CDC/NAS guidelines (proportional allocation); within-state allocation and the impact of vaccine hesitancy are left to the discretion of each team
  • Vaccine immunity delay: There is approximately a 14 day delay according to the Pfizer data; because we suspect the post first dose and post second dose delays may be of similar length, we do not believe there is any need to explicitly model a delay, instead groups can delay vaccine receipt by 14 days to account for it
  • Vaccine uptake: See specific details.
  • Vaccine rollout: rollout to follow ACIP recommendations unless known to be contradicted by state recommendations
    • Phase 1a: health care workers, long-term care facilities
    • Phase 1b: frontline essential workers, adults 75+
    • Phase 1c: other essential workers, adults with high-risk conditions, adults 65-74
  • NPI assumptions: phased reductions of NPIs in 2021 that align with patterns observed at different times in the course of the epidemic in 2020-21 (see scenario-specific guidance); teams have some liberty on how to implement these reductions within the guidelines
  • Database tracking of NPIs: teams may use their own data if desired, otherwise we recommend the following sources as a common starting point:
  • Geographic scope: state-level and national projections
  • Results: some subset of the following
    • Weekly incident deaths
    • Weekly cumulative deaths since start of pandemic (use JHU CSSE for baseline)
    • Weekly incident reported cases
    • Weekly cumulative reported cases since start of pandemic (use JHU CSSE for baseline)
    • Weekly incident hospitalizations
    • Weekly cumulative hospitalizations since simulation start
    • Weeks will follow epi-weeks (Sun-Sat) dated by the last day of the week
  • “Ground Truth”: The same data sources as the forecast hub will be used to represent “true” cases, deaths and hospitalizations. Specifically, JHU CSSE data for cases and deaths and HHS data for hospitalization.
  • Metadata: We will require a brief meta-data form, TBD, from all teams.
  • Uncertainty: aligned with the Forecasting Hub we ask for 0.01, 0.025, 0.05, every 5% to 0.95, 0.975, and 0.99 quantiles
  • Ensemble Inclusion: at present time, in order to be included in the ensemble models need to provide a full set of quantiles

Common Specifications

NPI: In contrast to past scenarios, we do not specify different levels of non-pharmaceutical interventions (NPI) use here. The future level of NPIs are left at the discretion of the modeling teams and should be specified in the teams’ metadata.

Vaccination

  • Doses available:
    • 50M Moderna/Pfizer 1st doses available monthly, June-November 2021
    • J&J no longer available (after May 2021)
    • Supply has likely eclipsed demand at this stage. Number of doses are for reference and as a reminder to account for different VE by manufacturer, but no longer indicate number of doses administered. Distribution of doses by manufacturer and associated vaccine efficacy should fit within these dose bounds.
  • VE:
    • 50% and 90% against symptoms (Moderna/Pfizer 1st and 2nd dose; vs. B.1.1.7/B.1.617 and other variants circulating in the projection period). This is based on reports from the UK and Israel indicating decreased protection against new variants such as B117 and B1617 after 1st dose, and no substantial decrease after 2nd dose. These estimates should be used against current and future circulating strains during June-Nov; however higher estimates of VE can be used to account for prior circulating strains.
    • VE is defined here as vaccine effectiveness against symptomatic disease. Teams should make their own informed assumptions about effectiveness and impacts on other outcomes (e.g., infection, hospitalization, death). Data on VE studies of infection and symptomatic disease are included below.

B.1.617+ variant strain with increased transmissibility: The scenarios define the B.1.617-like variants as 20% and 60% more transmissible than B.1.1.7 and other strains circulating in the US and is at 5% national prevalence on May 29, 2021. This 5% proportion on May 29th is a national estimate; teams can use state-specific data if they wish to. Timeframe of the increase in variant prevalence is up to each team, but it should be assumed the variant(s) become dominant due to increased transmissibility. The variant is more transmissible but it is not an immune escape variant; further, no decline of immunity from vaccination (other than VE) or natural infection should be modeled for B.1.617+ or other circulating variants. Other assumptions are at the discretion of each team, but should be documented in metadata. More info on next page.

Vaccination Hesitancy

Vaccine hesitancy is expected to cause vaccination coverage to slow and eventually saturate at some level below 100%. The saturation levels provided in these scenarios are illustrative National reference points to guide defining hesitancy. The high vaccination scenario (low hesitancy) saturates at 86% vaccination coverage nationally among the vaccine-eligible population (updated from 83% in Round 5), as defined by current estimates from the Delphi group (link) (red line in figure, borrowed from round 5, but the same spirit applies to round 6). The low vaccination scenario (high hesitancy) saturates at 75% vaccination coverage nationally among the vaccine-eligible population, defined by the lowest county-level estimate from the U.S. Census Bureau’s Pulse Survey (link) from April 24, 2021 data. The speed of vaccination saturation should be defined by the modeling teams, and can be defined as a logistic function (red and blue lines in figure below) or at different speeds (green line below). State or smaller geospatial unit hesitancy limits should be defined by the modeling team using their best judgment. Overall national hesitancy should be similar to the illustrative levels defined in the scenarios, though is not expected to be exact. The eligible population for vaccination is presumed to be individuals aged 12 years and older.

Submission Information

Scenario Scenario name for submission file Scenario ID for submission file
Scenario A. High Vaccination, Low Variant Transmissibility Increase highVac_lowVar A-2021-06-08
Scenario B. High Vaccination, High Variant Transmissibility Increase highVac_highVar B-2021-06-08
Scenario C. Low Vaccination, Low Variant Transmissibility Increase lowVac_lowVar C-2021-06-08
Scenario D. Low Vaccination, High Variant Transmissibility Increase lowVac_highVar D-2021-06-08
  • Due date: June 8, 2021
  • End date for fitting data: May 29, 2021 (no fitting should be done to data from after this date)
  • Start date for scenarios: May 30, 2021 (first date of simulated transmission/outcomes)
  • Simulation end date: November 27, 2021 (26-week horizon)

Scenario and Simulation Details

  • Social Distancing Measures:
    • Includes combined effectiveness/impact of all non-pharmaceutical interventions and behavior change.
    • Current and future levels of social distancing are to be defined by the teams based on their understanding of current and planned control and behavior and expectations. Teams should consider that most jurisdictions are opening fairly quickly. No reactive interventions should be planned.
  • Testing-Trace-Isolate: constant at baseline levels
  • Masking: Included as part of “Social Distancing Measures” above.
  • Vaccination:
    • Pfizer / Moderna
      • Vaccine efficacy (2-dose vaccines):
        • B.1.1.7, B.1.617+, and currently circulating variants in the US
          • First dose: 50% against symptoms, 14 days after 1st dose
          • Second dose: 90% against symptoms, 14 days after 2nd dose
        • Effectiveness and impact on infection and other outcomes (hospitalizations, deaths) is at team’s discretion and should be clearly documented in team’s metadata.
        • Doses 3.5 weeks apart
      • Vaccine availability:
        • December-May: based on data on administered doses
        • June-November: 50 million available first doses/month, with the intention of protocols being followed (100M doses/mo)
    • Johnson & Johnson
      • Vaccine efficacy (1-dose):
        • 70% VE against previous strains; 60% VE against B.1.1.7/B.1.617+
      • Vaccine availability:
        • March-May: based on data on administered doses, with continuing at rate current on date of projection for remainder of month (~10M total administered).
        • June-November: No longer available; only 10M of 20M doses administered, supply, safety, and demand issues.
        • Manner for accounting for protection provided in the 10M vaccinated during March-May at team's discretion.
  • Vaccine Hesitancy: Vaccine hesitancy expected to cause vaccination coverage to slow and saturate below 100%. National vaccination saturation levels designated for each scenario serve as illustrative reference points to guide defining hesitancy, though the speed of that saturation and heterogeneity between states (or other geospatial scale) and/or age groups are at the discretion of the team.
  • B.1.1.7 Variant strain: Teams should model the B.1.1.7 variant as appropriate to their model. Any assumptions (e.g., differences in severity/mortality, VE, or natural immunity) should be clearly defined in the metadata. The default assumptions are that the variant is 1.5x more transmissible than current strains and reaches 50% dominance by March 15 and 100% dominance by May 1 (see MMWR report); here a 1.5x increase in transmissibility is defined as the increase in the expected number of infections by a single infected individual over their entire course of infection when there are no interventions or immunity in the population (e.g., a 1.5x increase in R0 in a classic epidemic model). No differences between B.1.1.7 and current strains in severity, mortality, or VE are assumed in default.
  • B.1.617+ Variant strain: Transmission advantage of B.1.617+ should follow the scenario guidelines; additional features are at the discretion of the team. Any additional assumptions (e.g., differences in severity/mortality, VE) should be clearly defined in the metadata. The scenarios define the B.1.617-like variants as 20% or 60% more transmissible than B.1.1.7 and other strains circulating in the US and is at 5% national prevalence on May 29, 2021. The 5% proportion on May 29th is a national estimate; teams can use state-specific data if they wish to. Here a 20%/60% increase in transmissibility is defined as the increase in the expected number of infections by a single infected individual over their entire course of infection when there are no interventions or immunity in the population (e.g., a 20/60% increase in R0 in a classic epidemic model). Timeframe of the increase in variant prevalence is up to each team, but it should be assumed the variant(s) become dominant due to increased transmissibility. No immune escape feature for B.1.617+.
  • Transmission assumptions: models fit to US state-specific dynamic up until End date for fitting data specified above – no proscribed R0, interventions, etc.
  • Pathogenicity assumptions: no exogenous fluctuations in pathogenicity/transmissibility beyond seasonality effects
  • Vaccine effectiveness: level of effectiveness and available doses are specified for each scenario; assumptions regarding time required to develop immunity, age-related variation in effectiveness, duration of immunity, and additional effects of the vaccine on transmission are left to the discretion of each team
  • Vaccine allocation: between-state allocation is based on population per the CDC/NAS guidelines (proportional allocation); within-state allocation and the impact of vaccine hesitancy are left to the discretion of each team
  • Vaccine immunity delay: There is approximately a 14 day delay according to the Pfizer data; because we suspect the post first dose and post second dose delays may be of similar length, we do not believe there is any need to explicitly model a delay, instead groups can delay vaccine receipt by 14 days to account for it
  • Vaccine uptake: See specific details.
  • Vaccine roll-out: roll-out to follow ACIP recommendations unless known to be contradicted by state recommendations
    • Phase 1a: health care workers, long-term care facilities
    • Phase 1b: frontline essential workers, adults 75+
    • Phase 1c: other essential workers, adults with high-risk conditions, adults 65-74
  • NPI assumptions: NPI estimates should be based on current trends and reported planned changes.
  • Database tracking of NPIs: teams may use their own data if desired, otherwise we recommend the following sources as a common starting point:
  • Geographic scope: state-level and national projections
  • Results: some subset of the following
    • Weekly incident deaths
    • Weekly cumulative deaths since start of pandemic (use JHU CSSE for baseline)
    • Weekly incident reported cases
    • Weekly cumulative reported cases since start of pandemic (use JHU CSSE for baseline)
    • Weekly incident hospitalizations
    • Weekly cumulative hospitalizations since simulation start
    • Weeks will follow epi-weeks (Sun-Sat) dated by the last day of the week
  • “Ground Truth”: The same data sources as the forecast hub will be used to represent “true” cases, deaths and hospitalizations. Specifically, JHU CSSE data for cases and deaths and HHS data for hospitalization.
  • Metadata: We will require a brief meta-data form, TBD, from all teams.
  • Uncertainty: aligned with the Forecasting Hub we ask for 0.01, 0.025, 0.05, every 5% to 0.95, 0.975, and 0.99 quantiles
  • Ensemble Inclusion: at present time, in order to be included in the ensemble models need to provide a full set of quantiles

Common Specifications

NPI: In contrast to past scenarios, we don’t specify different levels of non-pharmaceutical interventions (NPI) use; however, teams should consider that most schools intend to return to in-person education in the fall. The future level of NPIs are left at the discretion of the modeling teams and should be specified in the teams’ metadata.

Vaccination

  • Doses available:
    • 50M Moderna/Pfizer 1st doses available monthly, June 2021-January 2022
    • J&J no longer available (after May 2021)
    • Supply has likely eclipsed demand at this stage. Number of doses are for reference and as a reminder to account for different VE by manufacturer, but no longer indicate number of doses administered. Distribution of doses by manufacturer and associated vaccine efficacy should fit within these dose bounds.
  • VE:
    • Optimistic (50% and 90% against symptoms (Moderna/Pfizer 1st and 2nd dose) vs. Delta) is based on reports from the UK and the manufacturers indicating decreased protection against new variants such as Alpha and Delta after 1st dose, and no substantial decrease after 2nd dose.
    • Pessimistic (35% and 85% against symptoms (Moderna/Pfizer 1st and 2nd dose) vs. Delta) is based on reports from the UK and Israel indicating further decreased protection against the Delta variant after 1st dose, and notable decrease after 2nd dose.
    • VE is defined here as vaccine effectiveness against symptomatic disease. Teams should make their own informed assumptions about effectiveness and impacts on other outcomes (e.g., infection, hospitalization, death).

Delta variant strain with increased transmissibility: The scenarios define the Delta (B.1.617.2) variant as 40% and 60% more transmissible than Alpha (B.1.1.7.) Initial prevalence should be estimated or defined by the teams based on sequencing and other relevant data, preferably at the state level. Timeframe of the increase in variant prevalence is up to each team, but it should be assumed the variant(s) become dominant due to increased transmissibility. The variant is more transmissible but it is not an immune escape variant; further, no decline of immunity from vaccination (other than VE) or natural infection should be modeled for Delta or other circulating variants. Other assumptions are at the discretion of each team, but should be documented in metadata. More info on next page.

Vaccination Hesitancy

Vaccine hesitancy is expected to cause vaccination coverage to slow and eventually saturate at some level below 100%. The saturation levels provided in these scenarios are illustrative National reference points to guide defining hesitancy. The high vaccination scenario (low hesitancy) saturates at 86% vaccination coverage nationally among the vaccine-eligible population (updated from 83% in Round 5), as defined by current estimates from the Delphi group (link) (red line in figure, borrowed from round 5, but the same spirit applies to round 6). The low vaccination scenario (high hesitancy) saturates at 75% vaccination coverage nationally among the vaccine-eligible population, defined by the lowest county-level estimate from the U.S. Census Bureau’s Pulse Survey (link) from April 24, 2021 data. The speed of vaccination saturation should be defined by the modeling teams, and can be defined as a logistic function (red and blue lines in figure below) or at different speeds (green line below). State or smaller geospatial unit hesitancy limits should be defined by the modeling team using their best judgment. Overall national hesitancy should be similar to the illustrative levels defined in the scenarios, though is not expected to be exact. The eligible population for vaccination is presumed to be individuals aged 12 years and older.

Submission Information

Scenario Scenario name for submission file Scenario ID for submission file
Scenario A. High Vaccination, Low Variant Transmissibility Increase highVac_lowVar A-2021-07-13
Scenario B. High Vaccination, High Variant Transmissibility Increase highVac_highVar B-2021-07-13
Scenario C. Low Vaccination, Low Variant Transmissibility Increase lowVac_lowVar C-2021-07-13
Scenario D. Low Vaccination, High Variant Transmissibility Increase lowVac_highVar D-2021-07-13
  • Due date: July 13, 2021
  • End date for fitting data: July 03, 2021 (no fitting should be done to data from after this date)
  • Start date for scenarios: July 04, 2021 (first date of simulated transmission/outcomes)
  • Simulation end date: January 01, 2022 (26-week horizon)

Scenario and Simulation Details

  • Social Distancing Measures:
    • Includes combined effectiveness/impact of all non-pharmaceutical interventions and behavior change.
    • Current and future levels of social distancing are to be defined by the teams based on their understanding of current and planned control and behavior and expectations. Teams should consider that most jurisdictions are opening fairly quickly, and most schools intend to return to in-person education in the fall. No reactive interventions should be planned.
  • Testing-Trace-Isolate: constant at baseline levels
  • Masking: Included as part of “Social Distancing Measures” above.
  • Immune waning and Immune escape: As defined by the modeling team.
  • Vaccination:
    • Pfizer / Moderna
      • Vaccine efficacy (2-dose vaccines):
        • B.1.1.7
          • First dose: 50% against symptoms, 14 days after 1st dose
          • Second dose: 90% against symptoms, 14 days after 2nd dose
        • B.1.617.2
          • First dose: 35% vs 50% against symptoms, 14 days after 1st dose
          • Second dose: 85% vs 90% against symptoms, 14 days after 2nd dose
        • Effectiveness and impact on infection and other outcomes (hospitalizations, deaths) is at team’s discretion and should be clearly documented in team’s metadata.
        • Doses 3.5 weeks apart
      • Vaccine availability:
        • December-June: based on data on administered doses
        • July-January: 50 million available first doses/month, with the intention of protocols being followed (100M doses/mo)
    • Johnson & Johnson
      • Vaccine efficacy (1-dose):
        • 70% VE against previous strains; 60% VE against B.1.1.7/B.1.617.2
      • Vaccine availability:
        • March-May: based on data on administered doses, with continuing at rate current on date of projection for remainder of month (~10M total administered).
        • June-January: No longer available; only 10M of 20M doses administered, supply, safety, and demand issues.
        • Manner for accounting for protection provided in the 10M vaccinated during March-May at team's discretion.
  • Vaccine Hesitancy: Vaccine hesitancy expected to cause vaccination coverage to slow and saturate below 100%. National vaccination saturation levels designated for each scenario serve as illustrative reference points to guide defining hesitancy, though the speed of that saturation and heterogeneity between states (or other geospatial scale) and/or age groups are at the discretion of the team.
  • Alpha (B.1.1.7) variant strain: Teams should model the B.1.1.7 variant as appropriate to their model. Any assumptions (e.g., differences in severity/mortality, VE, or natural immunity) should be clearly defined in the metadata. The default assumptions are that the variant is 1.5x more transmissible than wild-type strains and followed the trajectory outlined here (see MMWR report); here a 1.5x increase in transmissibility is defined as the increase in the expected number of infections by a single infected individual over their entire course of infection when there are no interventions or immunity in the population (e.g., a 1.5x increase in R0 in a classic epidemic model). No differences between B.1.1.7 and wild strains in severity, mortality, or VE are assumed in default.
  • Delta (B.1.617.2) variant strain: Transmission advantage of Delta should follow the scenario guidelines; additional features are at the discretion of the team. Any additional assumptions (e.g., differences in severity/mortality, VE) should be clearly defined in the metadata. The scenarios define the Delta variant as 40% or 60% more transmissible than Alpha and other strains circulating in the US. Prevalence is not pre-specified – teams are expected to define this on their own. Here a 40%/60% increase in transmissibility is defined as the increase in the expected number of infections by a single infected individual over their entire course of infection when there are no interventions or immunity in the population (e.g., a 40/60% increase in R0 in a classic epidemic model). Timeframe of the increase in variant prevalence is up to each team, but it should be assumed the variant(s) become dominant due to increased transmissibility. No immune escape feature for Delta variant.
  • Transmission assumptions: models fit to US state-specific dynamic up until End date for fitting data specified above – no proscribed R0, interventions, etc.
  • Pathogenicity assumptions: no exogenous fluctuations in pathogenicity/transmissibility beyond seasonality effects
  • Vaccine effectiveness: level of effectiveness and available doses are specified for each scenario; assumptions regarding time required to develop immunity, age-related variation in effectiveness, duration of immunity, and additional effects of the vaccine on transmission are left to the discretion of each team
  • Vaccine allocation: between-state allocation is based on population per the CDC/NAS guidelines (proportional allocation); within-state allocation and the impact of vaccine hesitancy are left to the discretion of each team
  • Vaccine immunity delay: There is approximately a 14 day delay according to the Pfizer data; because we suspect the post first dose and post second dose delays may be of similar length, we do not believe there is any need to explicitly model a delay, instead groups can delay vaccine receipt by 14 days to account for it
  • Vaccine uptake: See specific details.
  • Vaccine roll-out: roll-out to follow ACIP recommendations unless known to be contradicted by state recommendations
    • Phase 1a: health care workers, long-term care facilities
    • Phase 1b: frontline essential workers, adults 75+
    • Phase 1c: other essential workers, adults with high-risk conditions, adults 65-74
  • NPI assumptions: NPI estimates should be based on current trends and reported planned changes.
  • Database tracking of NPIs: teams may use their own data if desired, otherwise we recommend the following sources as a common starting point:
  • Geographic scope: state-level and national projections
  • Results: some subset of the following
    • Weekly incident deaths
    • Weekly cumulative deaths since start of pandemic (use JHU CSSE for baseline)
    • Weekly incident reported cases
    • Weekly cumulative reported cases since start of pandemic (use JHU CSSE for baseline)
    • Weekly incident hospitalizations
    • Weekly cumulative hospitalizations since simulation start
    • Weeks will follow epi-weeks (Sun-Sat) dated by the last day of the week
  • “Ground Truth”: The same data sources as the forecast hub will be used to represent “true” cases, deaths and hospitalizations. Specifically, JHU CSSE data for cases and deaths and HHS data for hospitalization.
  • Metadata: We will require a brief meta-data form, TBD, from all teams.
  • Uncertainty: aligned with the Forecasting Hub we ask for 0.01, 0.025, 0.05, every 5% to 0.95, 0.975, and 0.99 quantiles
  • Ensemble Inclusion: at present time, in order to be included in the ensemble models need to provide a full set of quantiles