How simulations work
There are three basic steps to run a simulation-based projection in CHOS.
- Step 1: Select state and county
- Step 2: Review or customize input data
- Step 3: Review simulation-based projection
Step 1: Select state and county
The COVID-19 Hospital Occupancy Simulator (CHOS) provides hospitals with a simulation-based projection supported with state-of-the-art forecasting models and the latest county-level data on the reported positive cases and metrics around COVID-19 hospitalizations. A simulation-based projection includes:
A hospital-level four-week projection that includes the following metrics for COVID-19 patients:
- Daily expected non-ICU occupancy and admissions
- Daily expected ICU occupancy and admissions net change
A county-level four-week projection of expected COVID-19 cases used to generate the hospital-level four-week projection
Therefore, when preparing a simulation for the first time, hospitals need to specify their county and the county's state.
- To learn how to start a simulation, see Start Simulation.
- To learn how the county-level four-week projection of expected COVID-19 cases is generated, see CDC County-Level Ensemble Model.
- The state-of-the-art forecasting models rely on certain assumptions to effectively generate a simulation-based projection. To learn more, see Modeling Assumptions.
Step 2: Review or customize input data
To generate a hospital-level four-week projection, the models use certain variables that obtain an average value by computing county-level data from the past four weeks around hospitalization metrics. At times, it could be the case that a hospital might not agree with the average values because the hospital knows to have above or below-average values (e.g., big vs. small hospitals in the same county). This perceived difference will often occur among big and small hospitals when county-level averages will not correspond. Therefore:
- Before starting a simulation, hospitals can modify the average values given to the model's variables through data overrides known as personalized hospital overrides. Through these overrides, CHOS enables hospitals to use their perceived average values while considering the specific characteristics of the hospital. To learn about the available personalized hospital overrides, see Overrides: personalized hospital.
On top of the personalized hospital overrides, hospitals can introduce certain overrides based on expectations about the future to play out different scenarios and generate a modified simulation-based projection. The data overrides that enable hospitals to modify future averages are referred to as future expert estimate overrides. To learn more, see Adjust simulation to reflect best- and worst-case alternatives.
- To learn about the particular county-level data points that are averaged to generate a hospital-level four-week projection, see Data Files.
Step 3: Review simulation-based projection
After the simulation-based projection is generated, the following interactive and simple charts will appear:
For the hospital-level four-week projection, the following charts are available:
- Occupancy and admissions, non-ICU
- Occupancy and net change, ICU
- Length of stay, non-ICU
- Length of stay, ICU
For the county-level four-week projection used to generate the hospital-level four-week projection, the following chart is available:
After a simulation-based projection is generated, hospitals can introduce certain overrides based on expectations about the future to play out different scenarios and generate an alternative simulation-based projection. The data overrides that enable hospitals to modify future averages are referred to as future expert estimate overrides. To learn more, see Adjust simulation to reflect best- and worst-case alternatives.
- To learn more about the charts and their interpretation, see Charts.
Adjust simulation to reflect best- and worst-case alternatives
Using one of the future expert estimate overrides enables hospitals to change a four-week simulation-based projection while taking into consideration the used future expert estimate overrides.
One of the available overrides enables hospitals to select a hospital-level average of future expected COVID-19 cases per week. This override allows hospitals to clearly understand the implications of an increasing or decreasing surge of COVID-19 cases within their facilities.
To learn about the available future expert estimate overrides, see Overrides: future expert estimates.
- Submit and view feedback for this page
- Send feedback about H2O Health to cloud-feedback@h2o.ai