States aren’t testing uniformly for coronavirus. That’s creating a distorted picture of the outbreak.
Some states are better off than others.

The United States has surpassed the rest of the world in the number of confirmed cases of COVID-19, with 85,762, as of today (March 27). That came as no surprise to most experts, as it took seven weeks after the first case of the disease was identified in the U.S. for the country to start testing en masse — plenty of time for the SARS-CoV-2 virus to spread, undetected.
Now, 10 weeks after the country's first case, testing in the U.S. is beginning to ramp up significantly, but not uniformly. As of Thursday (March 26), 20 states were only up to a testing rate of 1 per 1,000 people. And six states had tested fewer than 1,000 people. That means it's difficult to know, based on reported numbers alone, how widely the coronavirus is circulating in a given state or community.
"The problem is we haven't had enough tests," said Steffanie Strathdee, the associate dean of Global Health Sciences at the University of California San Diego's Department of Medicine and author of "The Perfect Predator" (Hachette Books, 2019). "There's not even enough tests to test the people who are sick."
Nearly half of all known U.S. coronavirus cases are in New York, which has reported more than 38,977 cases. As of March 26, the state had recorded 103,479 tests, according to The COVID Tracking Project, a nationwide dataset managed by volunteer analysts and journalists. With a population of 19.4 million, that's a testing rate of around 5 tests per 1,000 people, which is the highest testing rate in the country.
Washington state is a distant second, having recorded 34,292 tests for the state's population of about 7.8 million. That's a rate of 4 tests per 1,000 people. In third place is New Mexico, which has run only about 7,800 tests, but that's a rate of 4 tests per 1,000 people (the state's population is just over 2 million).
The two most populous states in the U.S. are California and Texas, with 39 million and 29 million people, respectively. So far, California is testing at a rate of 2 tests per 1,000 residents, but about 57,400 tests are still waiting on results, according to the state's Department of Public Health. Texas is testing at a rate of only 1 test per 1,000 people.
Why is there such a big difference between states?
Testing for SARS-CoV-2 got off to a rough start in the U.S. On Feb. 5, the Centers for Disease Control and Prevention sent diagnostic kits for SARS-CoV-2 to about 100 public-health laboratories across the country. Most of the labs received faulty kits, which meant testing had to continue exclusively at CDC headquarters until the agency could develop and send out replacement kits.
On Feb. 29, the U.S. Food and Drug Administration commissioner announced that the agency would allow local public health and commercial labs to develop and use their own tests for the novel coronavirus. So, now we have a mix of public health labs and commercial labs testing across the country.
Some states have better public health laboratories than others, said Dr. Jeffrey Klausner, an epidemiologist at the University of California Los Angeles Fielding School of Public Health. New York, for example, has some of the best public health institutions in the country. But labs in other parts of the country are far less equipped. "Public-health laboratories were never really capacitated to be on the front lines of a large disease outbreak," Klausner told Live Science.
Here is the March 26 update of US per-state semi-log chart of case density and trends. NY, NJ, LA, and WA continue to have the highest case densities. There are now 14 states higher in case density than Italy on March 9th when their country-wide lock down occurred. pic.twitter.com/i40z3cFAqbMarch 27, 2020
Many states are depending on commercial laboratories to help run the tests, but so far, those labs seem unable to pivot from their normal testing routines, Klausner said. "There are huge delays in commercial labs," he said. "I have patients that have been waiting for over 10 days now. That's not helpful for patient care or public health."
Klausner and his colleagues in Los Angeles were so frustrated and tired of waiting on testing results in their area that they created a new commercial laboratory themselves in just eight days. They're now running over 500 tests a day, and have the capacity to scale up exponentially. "These kinds of high-capacity, high-throughput laboratories need to be replicated in high outbreak areas," he said.
The attitude and priority placed on testing also seems to differ between state governments, Strathdee said. This might be because there are mixed messages coming from the federal level. Not long ago, President Trump downplayed the threat of the novel coronavirus, assuring Americans it was nothing to worry about. "Governors might be using that as an excuse, and they might be in rural states that aren't yet seeing a huge deluge of cases," Strathdee told Live Science. "The 'not here, not us' attitude is prevalent."
Why is testing important?
"Testing and case identification is critical both for patient care and public health to control these local outbreaks," Klausner said. There are other ways to measure an outbreak, such as the number of doctor visits, emergency room visits and admission to the hospital, he said. That's enough to tell epidemiologists that the outbreaks are geographically limited. But the lack of testing has left states unable to implement a strategic approach to containing the outbreaks. "So politicians have had to resort to these massive statewide shutdowns that would not be happening if there was more testing capacity," he said.
In a perfect world, doctors find out who is infected early on, then isolate those patients and trace their contacts and test them, too. But the U.S. isn't anywhere close to being able to do that right now, Strathdee said. Other places, like South Korea and Canada, have done a much better job at this, she said. Health professionals in those countries can use what's called the ring method, where they essentially draw a ring around cases and test everyone inside the ring and figure out who needs to be isolated. Probably the closest the U.S. has come to this was in New Rochelle, New York, where about 1,000 people were asked to self-quarantine after they were traced back to a lawyer in the community who had COVID-19. It's still too soon to tell, but early data suggests that strategy blunted the disease spread, New York Gov. Andrew Cuomo said in his daily briefing on Wednesday (March 25).
In the U.S., testing data isn't being collated by a government agency. Instead, it's been up to volunteer citizens to create open-source platforms like The COVID-19 Tracker Project, to track the data. It's an amazing effort, Strathdee said, but nobody really knows how complete or accurate the data are.
What is clear, though, is that the number of known cases in the U.S. continues to grow at an exponential rate, with no sign of slowing down, she said. Having more testing data will help make decisions once infection rates have slowed and the ring method can be used effectively, but it probably won't help right now.
"There's something to be said for data but then there's observations on the ground where you can tell it's not going to turn around anytime soon," she said. "Why check the weather forecast when you can look outside and see there's a thunderstorm coming?"
Nonetheless, testing rates are now increasing nationwide, and that's promising news for everyone. "We're starting to see signs that people are mobilizing and that people are serious and we're all in this together," Strathdee said. "I'm heartened by that."
Originally published on Live Science.
- Going viral: 6 new findings about viruses
- Top 10 mysterious diseases
- The 12 deadliest viruses on Earth
The one-month trial gives you access to all of the educational site's 9,000 activities in reading, science, math and art. Keep your child busy and learning while we are all stuck indoors.
Live Science newsletter
Stay up to date on the latest science news by signing up for our Essentials newsletter.
Kimberly has a bachelor's degree in marine biology from Texas A&M University, a master's degree in biology from Southeastern Louisiana University and a graduate certificate in science communication from the University of California, Santa Cruz. She is a former reference editor for Live Science and Space.com. Her work has appeared in Inside Science, News from Science, the San Jose Mercury and others. Her favorite stories include those about animals and obscurities. A Texas native, Kim now lives in a California redwood forest.
-
chrisko Of course they're not. Is there really any reason to expect them to be? No country has yet tested everybody. So far, every single country has tested those who show symptoms or who have been exposed to others of known infection. This is true for S. Korea, China, Italy, Spain, USA - everywhere in the world. Everywhere in the world, there is a lack of tests available to test everybody - this is hardly unique to the USA (contrary to what some mediocre US media outlets seem to proclaim). As of late March, I know of only one "population" that has seen 100% population testing - the Diamond Princess Cruise Ship. Of the roughly 3,600-3,700 on board the ship, all were tested. The last I heard/read (a statistical medical study on the Diamond Princes Cruise ship's coronavirus history) was that roughly 800 had tested positive in the end (included post ship-departure quarantine periods) and 10-11 (updated as I had heard very recently that an 11th person had died - but it was not clear if that person contracted the virus from the ship or since departure of the ship and after the quarantine period) of that 800 had died. Now, that was a cruise ship and by age and health "condition" was composed of an exceptionally high-risk group of individuals (based on age and proclivity to pre-existing conditions stemming largely from the age group and cross referenced to death rates in China by age breakdown). The death rate of that high risk group has so far (mid to late March) shown to be just barely above 1%. In the end, I suspect based on this and "transferring" that data to the global population (or that of any modernized country at least), that the realistic death rate of infected people will be in the range of 0.5% to 1.5% (remember, the high-risk population of infected people from the Diamond Princess Cruise ship hovers right around 1% and we are more than a month since the ship has been emptied). In populations that have unusually high numbers of elderly people or communities that have a high level of "poor health" or "poor health habits", then perhaps for these populations - the death rates may range more in the 1% to 2% range. Like for Italy (which has the second oldest population in the world, has a healthcare system that was failing even before Covid-19 hit and has social family habits/lifestyles that promote much more intergenerational habits - vs. that which we encounter/practice here in the USA). So as a result, IMO, the best method to determine the number of infected people in a population (say New York City or New York State) is to look at the actual number of deaths and extrapolate from that number the range of the likely number of people infected. So if there are 1,000 confirmed deaths from Covid-19, we can then extrapolate that there are likely between 66,600 (assuming a 1.5% death rate) AND 200,000 (assuming a 0.5% death rate) people that are infected. So if there are 5,000 deaths in NY State - then we can just multiply/extrapolate the number of likely people infected - using this number it would range from 330,000 to 1,000,000 infected people (counting those who showed no signs or who have completely recovered). In the end, and as we increase (like every other country) the number of test kits available for testing over time, they will become less and less important over time in telling us important information - though the benefit will be to identify those who are infected, track their contacts and remove them from society (quarantine them). In the end, way too many people will 1: not be tested as they show zero signs AND/OR 2: be tested if the goal is to test 100% of the population, but will have already recovered. By mid summer, we will move from the key being testing of people for active infection to testing of people for past infection - and only then and over the coming 2 years (when people have their regular doctor's appointments) will we really know just how may American's were infected with Covid-19.Reply
In the end, we cannot easily stop pandemics. This idea that our government should be ready with huge stockpiles of "ventilators" (for example) for a what if scenario isn't really very wise. Say we stock pile a quarter million ventilators and the next pandemic turns out to be blood based! Depending on who is in office, the media will proclaim why didn't that administration stockpile devices that can remove, clean and re-insert that "cleaned" blood in a stockpile. Or if that is a preferred administration - the media will ask why didn't the prior (unliked) administration didn't prepare us for this obvious outbreak (and the media will find a one-off Ted Talks about some person suggesting the next pandemic would be blood based - not respiratory in nature - and then use that as proof that we should have seen it coming). No, we don't know whether the next pandemic will be blood, brain, nervous system or some other life threatening area of our bodies that will be attacked. So what should we do to be prepared? There are a few things! First and foremost is the recognition that contrary to some people's beliefs - the USA MUST continue to be a manufacturing country! And not just in finished products, but in the whole supply chain. That for the government to step in and support this, we should implement mandatory minimum requirements for certain industries to purchase a specified percentage of their product from American manufacturers who use a 100% American-based supply chain of components, ingredients, etc. . . For example, we can mandate that all US hospitals purchase no less than 15% of their medical masks and PPE items from American manufacturers using American supply chains. Secondly, instead of having the government/taxpayers invest huge quantities of money in stockpiling equipment that over the years will become obsolete, we instead have the government stockpile the means to insure that equipment or goods can be readily put into place to increase the manufacturing capacity in short order. For example, instead of stockpiling 2 billion masks, we stockpile and rotating 25 million masks. But we secure the equipment (molds, machines, automation equipment) that could be quickly installed and producing those products in large quantities. This could be in the molds needed to manufacture parts for ventilators, blood machines, or the equipment necessary to combat various different types of equipment that may be needed to address a future pandemic. This would be far more efficient, far more cost effective and far more likely to provide for the ability to quickly ramp up production of future unknown products or equipment to manufacture the needed items.
The reality is, we are actually far more likely to encounter a bio-terrorism event (though the likelihood is that it would be more prone to be regional or contained in a smaller area) than another catastrophic pandemic - and this in the end creates the same end result in the needed response. The question is how much are we willing to pay in advance? How much are we willing to pay for national security? Because that is what this is really about - a country that can address it's own needs from food to medicines to technology to manufacturing to supply chains is far more nationally secure than any country that must rely on an outside country to support and provide for it's needs. This requires us to throw away some of those feel-good PC philosophies that provide the grandiose belief that we are a better country for addressing the needs of other countries and people that are not American's - than we are for addressing our own needs and our own people first and foremost. It's much like when we fly and are told to put on our own oxygen masks before helping others. As the idea of helping others requires one to first be in the position to help themselves first! We need to protect our own industries (manufacturing, pharmaceutical, agricultural, energy, mining) to protect our own future - after which we will be better prepared to help others in their times of need.
Most Popular
By Harry Baker
By Sascha Pare
By Ben Turner