Learn more about why the positivity rates shown on our site may differ from state calculations.
Since we launched the COVID-19 Testing Insights Initiative on the Johns Hopkins Coronavirus Resource Center (CRC), both the testing landscape and the data environment in the United States have changed considerably. This page explains how those changes result in differences between our site’s calculations and some state estimates.
It is important to note that test positivity is a measure of testing capacity and while it can provide important context about case totals and trends, it is NOT a measure of how prevalent the virus is in communities. Policy decisions, like openings and closings or interstate travel, should not be determined based on test positivity alone.
In the U.S., there are no federal standards for reporting COVID-19 testing data. This makes it impossible to offer a fully apples-to-apples view of testing data at the national level. Without federal standards, states have been left to forge their own paths, and as a result, they report testing data differently.
Under the current conditions, inputs into the same data categories differ between states. For example, in one state, the data for the number of tests administered might include both antigen tests and PCR tests. In another state, the testing data might only include PCR tests. This means that while the data category (“number of tests”) is the same, the inputs and resulting calculation are different.
Since the beginning of the pandemic, states have changed the amount and the type of testing data they report, and have been inconsistent in how they report antigen tests.
Some states also periodically pause or fully stop sharing key data that are used in making positivity calculations, or change the cadence with which they report data. Both of these actions can create abnormal spikes in positivity rates in tracking efforts such as ours.
These inconsistencies may result in a test positivity calculation on our site that is different from what states report.
According to the CDC, and based on the testing data that states currently share, there are several possible ways to calculate test positivity. A state’s positivity rate will be different depending on which approach is used. Per the CDC, each of these approaches are valid, but they provide different insights into the impact and limitations of a state’s testing efforts.
We believe it is important to look at both people-centered calculations and test-focused calculations, as each provides different information about a state’s testing efforts. We have identified 4 possible ways to calculate positivity. Below, you can view the variation in states’ positivity rates based on each of the approaches, and the wide discrepancies in the availability of data needed to make each of these calculations:
The Coronavirus Resource Center’s current approach to calculating positivity throughout our site is Approach 4, for the following reasons:
The lack of federal standards creates significant inconsistencies in how states report testing data. Currently, Approach 4 is the only one that can be used for all 50 states.
Our data scientists and epidemiologists believe a people-centered calculation allows users to gauge whether states are casting a wide enough net with their testing to identify infections that may be occurring.
By looking at the percentage of people who test positive, we also can see whether there are testing participation, access, or capacity problems that need to be addressed.
It is worth noting that historically, Approach 4 focused more on the number of people tested as opposed to the number of tests given, and is the only people-centered calculation for which all states report the necessary data. We routinely monitor the data environment and make changes if necessary.
Depending on how states report data on the number of people tested (e.g., whether they deduplicate data for each person tested), calculating positivity using Approach 4 may not capture repeat tests performed on the same person.
The CRC believes that as the pandemic has progressed and testing capacity has expanded, there are important public health reasons to conduct multiple tests on the same person.
For example, someone who was exposed early in the pandemic may have a subsequent exposure and need to be retested.
Additionally, states are increasingly conducting routine, repeat screening tests in high-risk populations, such as those in nursing homes.
At the same time, including these repeat tests may obscure important trends and early warning signals. For example, there is an increasing amount of screening testing being conducted within states for certain populations – such as colleges, universities, and large employers.This means that at one point in Washington D.C., tests performed by a major university represented approximately 20 percent of all tests conducted, which could skew the data for the District broadly.
Despite the limitations outlined above, we believe it is important to continue to calculate and track each state’s test positivity using Approach 4, which is a people-centered calculation, where possible. We continue to make every effort to use the best data available and will continue to monitor the data environment and make changes as necessary.
To explore this issue further, read Dr. Jennifer Nuzzo's interview with The Baltimore Sun.
States are not consistent in how and when they release and update their data, and some may even retroactively change the numbers they report. This can affect the percentages you see presented in these data visualizations. We are taking steps to account for these irregularities in how we present the information, but it is important to understand the full context behind these data.