Creating COVID-19 tests is complicated science, and business

The United States faces an ongoing shortage of tests for COVID-19, the disease caused by the coronavirus. Testing availability varies widely by state. And in New York City, the hardest-hit area of the U.S. so far, doctors are only offering tests to people with severe symptoms. The initial test used in the U.S. was created by the federal Centers for Disease Control, and it wasn’t always accurate. It was also only allowed to be used by state-run health departments. 

Without testing, we can’t tell how many people might have the virus but not show symptoms. We can’t predict the next outbreak, or even trace how the virus is spreading. Scientists and biotech companies around the country are working furiously to create alternate tests. I spoke with one of those people, Jacqueline Linnes, a professor of biomedical engineering who runs a lab at Purdue University. She’s created a COVID-19 test. She told me the science is complicated, but getting from lab to market is even harder. The following is an edited version of our conversation.

Jacqueline Linnes: There’s a difference between getting a test to work in the lab 5, 10 or even 100 times, and then getting it to a point that’s manufacturable at scale that we need. A lot of times when these are in development, the exciting part is getting the actual sensing to work. There’s lots of really cool sensing techniques that are out there that actually are not deployable at all, so there’s a big disconnect between what is manufacturable at scale and what is actually getting made in the lab. For us, if we are developing a device and we’re able to publish a paper on it, we’re super-excited. It’s only after that, that we come back and look at it and realize, “This has way too many layers and alignment issues.” Getting that discussion with manufacturers in the process is really critical. In companies that are doing manufacturing, it still takes a lot to switch from what they’re currently developing to developing a whole new set of biology and reagents and then getting that switched on to their existing platform. Even that is difficult.

Molly Wood: Is this changing the development process, at all? Is it unusual, for example, for a university to be in conversation with a manufacturer at the same time as the research is happening?

Linnes: It shouldn’t be, but I think that, yes, it is unusual for this. It has certainly sped up things in my lab as far as developing technologies with manufacturing in mind. It’s been beneficial, I think, so that we can reframe how we’re developing these devices and focus on how we’re not just going to make a cool device, but how it’s actually going to go from the lab into usability.

Wood: You said it shouldn’t be. Do you think that should have been the norm all along?

Linnes: In an ideal world, it would be the normal all along. In academia, we tend to be thinking about not just what functions now but what can we make function in five or 10 years, and hope that the other side of the manufacturing will have caught up by then. That disconnect has been existing and we are rapidly realizing that is not how it should be.

Wood: Are you saying that the lack of that coordination from before the pandemic is contributing to a slower rollout of testing now?

Linnes: Despite a number of people that have argued that we should be doing this all along, most research labs have paid no mind and have continued working on things that are much sexier than manufacturing devices.

Wood: Got it. Is money a part of that? Has there not been funding for the not-very-sexy parts?

Linnes: Always. I think that when we’re submitting proposals, and they undergo peer review, people get really excited about the innovative new biosensor and they’re not very excited about the mundane development processes to get things out into the world. The way that we’re even reviewing each other is contributing to that.

Wood: Is it the province of universities and research labs to do that work in the first place? Could that work, on the boring stuff, have been happening at the government level, ideally?

Linnes: That is a great question. I think that valley of death is something that nobody’s quite figured out whose job it is. Companies are out there scouting for potential technologies, but the further along the universities can get them, the better, because it’s a risk for them to translate something out of the lab into the company. I know at our university, we have a role-to-role manufacturing facility that’s specifically designed to help make that transition. People are thinking about it and we’re still figuring out the best place for it.

Related links: More insight from Molly Wood

Like so many things about this pandemic, it’s hard to tell what is and what isn’t true about testing. Several companies have been promoting at-home testing kits, but then the FDA came out and basically said you shouldn’t use them. The at-home swabbing is super-difficult, easy to screw up, the samples can be compromised by human error and being sent through the mail can result in false negatives. 

There are reports of a test that takes just 15 minutes to give results, instead of the hours or days required by other tests. That’s apparently messy because it’s not sensitive enough and can only detect about 60% of cases.

There are a lot of people out there who think they might already have had COVID-19 and want to take a test to find out. Those tests are starting to roll out, but Business Insider notes we still don’t know if people can get reinfected or not. 

Scientific American has a good roundup of what’s available, where and how the most common tests work, if you want to read up on how biotechnicians scan a nasopharyngeal sample using a reverse transcription polymerase chain reaction. Listen, I know some of you have a lot more time these days.