The CDC guidelines propose the use of six “gating” indicators to assess when a state should relax or increase its restrictions. The indicators are based on symptoms, cases, and hospitals.

This is a work in progress. As of June 15th, we've calculated 5 of the 6 gating indicators. You can find the code we use on GitHub. If you have feedback, send it to feedback@covidexitstrategy.org.

The data presented here was pulled from The COVID Tracking Project on Friday, July 10th, 2020.

This is a work in progress. As of June 15th, we've calculated 5 of the 6 gating indicators. You can find the code we use on GitHub. If you have feedback, send it to feedback@covidexitstrategy.org.

The data presented here was pulled from The COVID Tracking Project on Friday, July 10th, 2020.

Indicators Based on Cases

To determine a downward trajectory, data are assessed using a smoothed curve. This differs from the typical approach of using a 3-day or 7-day rolling average.

To calculate this curve, the CDC applies a cubic spline, or “smoothed curve”, a statistical method that smooths out day-to-day variability in the data. The slope of this curve is used to assess declining incidence. Using guidance provided by the CDC, we've calculated the 3-day cubic spline (3DCS).

To calculate this curve, the CDC applies a cubic spline, or “smoothed curve”, a statistical method that smooths out day-to-day variability in the data. The slope of this curve is used to assess declining incidence. Using guidance provided by the CDC, we've calculated the 3-day cubic spline (3DCS).

Decreases in newly identified COVID cases

According to the guidance, states are allowed a 5 day grace period when calculating a downward trajectory. Near-zero incidents is defined as fewer than 10 cases per 100k population over 14 days.

The data powering this table can be found here.

According to the guidance, states are allowed a 2 to 3 day grace period when calculating a downward trajectory. A near-zero plateau was not defined, so we've assumed it means less than 1% positivity.

The data powering this table can be found here.

Indicators Based on Symptoms

Decreases in visits for influenza-like-illness (ILI)

Only weekly data for ILI is available from the CDC, so trajectory is calculated based on the difference between two weeks. The raw data from the CDC can be found here.

The data powering this table can be found here.

Indicators Based on Health System

Treat all patients without crisis care

This indicator also looks at staff shortages and PPE availability. Because that information is not available publicly, we focus on inpatient & ICU occupancy.

The data powering this table can be found here.

Robust testing program

This indicator also looks "median time from test order to result" but that information is not made publicly available. So we focus on percent positivity.

The data powering this table can be found here.