Outside of the original epicenter in the Northeast, COVID-19 cases are rising across the United States. In fourteen states, the number of new cases discovered each day has increased in the last two weeks, some quite sharply: in Georgia, cases are up 27 percent; in California 37 percent; in Florida, 55 percent; in Texas, 56 percent. In Arizona, cases have risen from 28,296 to 58,179, an increase of more than 105 percent.
But we also know that the United States has substantially expanded COVID-19 testing. Testing numbers we compile from states and territories are up 30 percent in the last two weeks alone, and the seven-day average for daily tests is finally over the 500k/day initial acceptable minimum from the Harvard Global Health Institute (which has now been increased to 900k/day).
Two charts that can help us understand the case count jumps
So how can we tell whether the increase in cases is an accurate indication of worsening outbreaks or an artifact of expanded testing? Rising COVID-19 hospitalizations are suggestive of worsening outbreaks, but if a state is doing enough tests to discover mild or pre-symptomatic cases, hospitalizations may lag behind positive tests by days or weeks.
There’s a simple way to get closer to a definitive answer about which case count numbers indicate new or worsening hotpots. The percent-positive rate shows us how many tests come back positive, out of all tests performed. This rate can help us interpret case count increases:
If case counts go up, but the percent-positive rate goes down or holds steady, the rise in cases can be partially explained by the increase in testing.
If case counts and the percent-positive rate both go up, the increase cannot be explained entirely by an increase in testing. If case counts and the percent-positive rate both go up, but testing decreases or holds steady, the rise in cases could indicate new outbreaks of the virus in communities.
In the states with big percentage increases in case counts, percent-positive rates tell a more specific story. California’s percent-positive rate has remained a steady 5 percent, and Georgia’s percent-positive rate is up from 7 to 9 percent—higher than anyone would like to see, but not a huge increase. Texas, on the other hand, has gone from 7 percent to 12 percent positive in the last two weeks, and Florida from 6 percent to 15 percent positive. Arizona remains an outlier among outliers: their already high 16 percent-positive rate has nearly doubled to 28 percent.
Another way to look at the relationship between case counts and test counts is to compare the rate of change for a state's case-count rise to the rate of change for its testing numbers. If we chart the week-over-week increase in cases and tests from the last two weeks, we can see more clearly that for a few states, increased testing really does seem to account for some or all of the increase in new cases. The same comparison also makes it clear that in the states with the worst case numbers, testing explains only a fraction of the rise in cases.
Testing is booming in Arizona, but even this major testing increase hasn’t caught up to the rise in new cases. Week over week, Texas has also posted a big gain in testing—but again, this growth has been outstripped by the increase in new cases. California, on the other hand, has increased testing by the same proportion as Texas, but has seen a much smaller increase in new cases—testing probably accounts for more of California’s rise in cases than Texas’s.
The starkest example, Florida, has actually slowed testing almost 10 percent, week over week, while seeing a 65 percent jump in the cases in the same time span. Florida’s rising cases, we can conclude, have nothing to do with expanded testing. South Carolina and Kentucky have followed the same pattern as Florida, with testing increases slowing while cases rise.
Given the high percent-positive rates and the degree to which case counts have outpaced tests in Arizona, Florida, and Texas, we interpret the case counts as indicators of sharply worsening outbreaks. In California and Georgia, where percent-positive rates remain relatively steady, the numbers are more difficult to interpret, but we’ll be watching them closely over the coming weeks.
Want the weekly insights from our data by email? Join our low-frequency email list.
Charts by Peter Walker of The COVID Tracking Project and COVIDCharts.tech.
Erin Kissane is a co-founder of the COVID Tracking Project, and the project’s managing editor.
Jessica Malaty Rivera has an MS in Emerging Infectious Diseases and is the Science Communication Lead at The COVID Tracking Project.
More “Testing Data” posts
How Probable Cases Changed Through the COVID-19 Pandemic
When analyzing COVID-19 data, confirmed case counts are obvious to study. But don’t overlook probable cases—and the varying, evolving ways that states have defined them.
20,000 Hours of Data Entry: Why We Didn’t Automate Our Data Collection
Looking back on a year of collecting COVID-19 data, here’s a summary of the tools we automated to make our data entry smoother and why we ultimately relied on manual data collection.
A Wrap-Up: The Five Major Metrics of COVID-19 Data
As The COVID Tracking Project comes to a close, here’s a summary of how states reported data on the five major COVID-19 metrics we tracked—tests, cases, deaths, hospitalizations, and recoveries—and how reporting complexities shaped the data.