Making sense of workplace virus testing

By Kevin J. Delaney

Rapid virus testing promises to make it possible to go back to work and school safely. But many of us still have questions—and the Trumps’ infection increased focus on virus tests’ limitations. For answers, I spoke with Dr. Caesar Djavaherian, co-founder and chief clinical innovation officer at Carbon Health. Dr. Djavaherian’s company is working with schools and organizations including Brookfield Asset Management, the Oakland A's, and Alaska Airlines on their testing and prevention plans. Here is a transcript, edited for clarity, of our conversation about the realities of testing in workplaces.

(You can read the highlights and more coverage in this week's edition of my Reset email newsletter. You can sign up to receive the newsletter for free each week here.)

Dr. Caesar Djavaherian | Courtesy Carbon Health

What’s the best practice for testing employees so that they can safely return to the office and continue working while minimizing the risks of outbreaks?

The real answer is that it depends on the scenario and what the work environment is like and what the risks are that each employee takes in coming to work. So, for example, we have employees who work on set for production companies where they're encountering many other people they're in close quarters with. And the lead actor's health might impact whether a production continues or not. So the testing cadence in those scenarios can be much more frequent than for someone who might be in the office once in awhile taking full protocols of distancing and masking while they're there—in which case we would recommend once a month or once a week, depending on what the risk is.

What happened with Trump and Melania testing positive really highlights what we've been promoting with our Covid Ready program all along, which is that no single intervention is enough to reduce the risk to close to zero. The single intervention that you know Trump is taking is testing. So people get tested multiple times, like daily, anyone who comes into his orbit, and that's not enough. And the reason is that someone can be infectious for three to four days before they test positive, but yet they're spewing virus to all those around them. So having multiple protocols in place is how we keep the workplace safe.

Despite having tens of thousands of employees who we've touched in one way or another, we haven't had any outbreaks. So people have brought the virus into the workplace, but haven't spread it to coworkers. And again, it's because we don't just do testing, but we do workplace protocols, prevention measures, education. So we're layering on top of one another each of these interventions.

You said one single intervention—even testing—is not enough. Why is that?

If you look at countries where the coronavirus pandemic has been successfully held at bay, there are countries that don't have any more technologies than we do. Taiwan, New Zealand in the early days, South Korea. What those countries did that the US didn't do is through very strong leadership. They were able to deploy all of the tools we have in the toolbox. We know the main tool is prevention, the education and prevention measures really explaining to people how the virus is transmitted. Like why masking is so critical, not just for your own health, but for the health of anyone we touch. And distancing measures. Then on top of that, you can layer testing to identify the asymptomatic patients as well as those symptomatic who test positive. Then if you have a positive case, it's the protocols around quarantine, contact tracing, making sure that those people around them are also isolated and then providing health care through telemedicine.

We don't rely on the average person when they wake up with a sore throat and say, ‘Oh, this is chronic virus and I'm staying home.’ Or ‘I always get a sore throat after I yell at the debates. And I'm just going to go to work and don't tell them.’ They don't have to make that judgment call because they just go on their app and request a visit within 10 minutes. So we are staffed such that within 10 minutes we can see patients anywhere in the country and provide real advice. Those are all the layers that we've put on top of our entire Covid Ready program to really create a safe workplace. That's really what the country should be doing as a whole. Certain countries are like Taiwan. We're doing it in the areas where we have influence, which is with corporations.

There are probably other takeaways from the Trump infections, but one of them is testing alone is not enough?

When my brother called me to tell me, my initial thought was that these guys confiscated like thousands and tens of thousands of rapid tests in the early days when Abbott first developed them so that they could test everyone in his orbit, including at whatever frequency they want. And yet he still got the virus. There's no clearer message that testing alone is not the solution. Saying ‘Ha, I wear a mask but only when I want to’—the real takeaway is you didn't use all the tools in your toolbox. You're relying on one of them, which is imperfect.

What tests should companies use?

There are really only three categories of tests. One category is PCR. The second category is antigen. An antigen is a protein created by the virus. The third category is antibody, your body's response to the virus proteins. The antibody test is useless for individuals and companies making a decision about whether a person is infectious or not. The antigen test is useful, but the sensitivity and specificity of the ones on the market and available today are a bit worse than the PCR tests. Under the PCR umbrella is saliva. The Abbott rapid test is also a PCR test. And commercial labs like LabCorp, Quest, and some of the genomic labs like Color and the Broad Institute are all PCR tests and that's kind of the gold standard right now. It's also an imperfect test because there's this dynamic between when you have enough virus in your system that the test actually detects it. 

But whether you're submitting a saliva sample or having a painful nasal swab or the not-painful kind of midturbinate swab or an oral swab, the test performance for all of the PCR tests is roughly the same. So we choose the one that's the fastest, because where we saw Major League Baseball fail is that they were relying on a saliva-based test that they're doing every other day for all their players. Well, we know that it takes three or four days for a test to turn positive, and then you're waiting an extra day or two for the results of that test to come back. You're really talking about one week or six days worth of time where someone can unknowingly infect others. So just by taking away the turnaround time with the rapid test, we can reduce the risk—not perfectly but dramatically. Since later on you're more infectious. So on day four or five or six, you're more infectious than you are on day one.

What's the fastest test you choose?

In the PCR world, it's the Abbott rapid test. That's a 15-minute test. In the antigen world, there's a company called Quidel and the platform is called the Sofia SARS test. That's also a 15-minute test, but it's an antigen test, so not awesome. But now Abbott has another antigen test that's supposed to be a $5, 15-minute test. The federal government has procured the first 30 million kits. So the average individual doesn't have access to them. We are working directly with Abbott to expand their indication—right now it's only allowed to be used in people who have symptoms. But we're working with Abbott to help them expand to asymptomatic screening. We're hoping to get that access to that in first quarter of 2021.

The Abbott rapid tests require a professional to administer them, right?

Yes. Actually all of them do.

What do they cost?

The price depends on the test type and quantity, and ranges from about $120 to about $180. Most of our customers are large enough to be at the $120 mark though.

The most-expensive tests are the ones where we send them to the laboratory because there's lab infrastructure and people. When you take into account labor costs, plus the tests themselves, it costs about the same, whether you're talking about the antigen test or the Quest, LabCorp, or the Abbott rapid tests. 

I think what's exciting is that the new Abbott antigen test is promising to be a $5 test, which would bring down overall test costs from roughly $120, when you're taking into account labor and such, down to probably $50 or $60. The lower the test costs, the better. We should then reach so many more people. So for the cost, we're all just kind of putting our eggs in the Abbott basket right now.

You said that it varies across types of workplaces, but is there a way to talk about best practices for a standard office?

For the average office worker, maybe in the financial industry or in tech, what we generally recommend is access to all of the education material around prevention, distancing, masking, all that stuff. We then ask these employees to enroll in our daily symptom tracker, which is just every day in the morning, you get a reminder to answer four or five questions. Our questions are not just questions that come from the CDC or the WHO, but we now have 300,000 people who we've tested through Carbon Health. And because we're a tech company we have really granular information about who tests positive. So what does that person look like? As well as what are their symptoms when they test positive versus testing negative. 

It turns out that headache is highly correlated with testing positive—not by itself, but if someone has a new headache and body ache, as an example, it's highly correlated with testing positive. We have that as part of our questionnaire—not just fever, cough, shortness of breath, things that in the early days we thought were important but turn out that they're less predictive than the things that we're asking you about today. So we ask everyone to enroll in this daily symptom tracker so that we can risk stratify based on their symptoms. And also, so we can provide really good healthcare information. Because if someone has diabetes and hypertension, we may be more conservative in our recommendations than if they're just a young, healthy person, a 20 year old. 

Then we recommend a baseline test for everyone. And that's really to identify those who the symptom tracker may not pick up on, since we know it's about 30% of [infected] people who will be asymptomatic—this baseline test is aimed at those folks. From there, we recommend a cadence of testing that ranges from once a week to once a month, looking for those asymptomatic individuals. 

In the meantime we make recommendations [for the workplace] around not using common areas, to having conference rooms where chairs are distanced by six feet, to having universal masking when people are inside the building, or frankly, outside the building. We also make sure that any employee has access to a provider via telemedicine on our platform whenever they have a question. When someone tests positive, we have one of our providers, either a nurse practitioner, a physician assistant, physician or a nurse reach out to that positive person and then really work backwards to see who they may have come into high-risk contact with, which is unmasked, more than 15 minutes, and all of the things that you're probably well aware of, so that we can then isolate those individuals and get them tested.

We might send them a saliva test so that they're not just waiting for the next time the test is available, but we're proactively reaching out to those who might be infected. This all works to keep the businesses going. I remember we had a construction contractor who tested positive, and the question is, do we close down our offices? Because this individual was in our space for multiple days. And the answer is absolutely not—the reason we have all these protocols is because we're actually assuming that someone in your organization is positive and is with you maybe sitting six feet away. We approach the problem with the assumption that the virus will be in your space.

That's interesting because a lot of the discussion in businesses centers around when someone catches it, do you shut down your workplace or send everyone home who was in the same floor as them.

With tens of tens of thousands of employees in our platform, every day it's like a new one-off scenario. The easy ones are: my coworker was tested positive, now what? But it's like ‘My college son came home and I gave him a hug. And then he ended up having symptoms. Do we shut down the business?’ Because the employee has been going to work since then. The answer is usually no, if all of these other protocols are in place. Our staff goes through and confirms that the at-risk individual hasn't put others at risk as well. 

We haven't had to shut down any businesses. We haven't recommended that. We have recommended that individuals stay home and be retested before allowing them to go back to work. But that's actually been rare because most of these companies are pretty strict about adhering to the workplace policies. That's our value add. It's like, you know, if someone tested positive or has symptoms and you shut everything down, then what good is Carbon? Like, we're not doing anything useful. You know, a lay person can make that decision.

I would have assumed, if I sat six feet away from someone who was infected, I would be sent home and everyone in the vicinity would be.

No, you definitely wouldn't. And in fact, our data support this approach. The best part of being a technology company is that it's not just the clinician’s gut instinct, which might differ from doctor to doctor. We have a team of data scientists who are constantly surfacing these interesting pieces for us. Whether it's proximity, symptoms are the easy ones where we have this massive data set, even things like language and socioeconomic background, we have all those data that we're looking at making our recommendations. So the answer is no. The reason I can say that confidently is because our data set is so large now. And we're fortunate enough to then convince the employees—because then it becomes a psychological thing, right?

Like, ‘Oh my God, the person sitting right across from you; sure he was six feet away. But I remember him sneezing and coughing a few times. I should stay home because I have an elderly parent that I live with.’ That's a real scenario that happened. So what do you say to that? So now it goes into the psychological component and then we'll get on the phone and or a Zoom meeting and really explain why they're safer at work than they are going to the grocery store. 

As an ER, doctor, I've been treating patients with coronavirus since February. Our hospital, North Bay Medical Center in Fairfield, had the first community transmitted case of coronavirus. My colleagues who intubated the patient, all the docs who took care of these people before, even knowing that they had coronavirus, very few of them got infected. And then we adopted workplace protocols so that now, even though the numbers have increased dramatically, we don't have anyone on staff who's gotten the virus from a patient. And we go into closed rooms with patients who are sick. So if doctors can remain safe, post that initial surge, I think companies can too. But they have to be really thoughtful in their approach.

Presumably in that specific scenario, it’s because you're wearing masks and keeping distance.

That's right. Masking, gloving, sanitization, distancing. We also walk into every room with the assumption that they could have coronavirus.

You're saying that the accuracy for the rapid test is similar to the test that people send out to labs, right?

Yes. So when the rapid test came out, the official FDA EUA [emergency use authorization] suggested that it was exactly on par. So it was not inferior to LabCorp and Quest. Maybe a month later, NYU Langone published an article that said hold on, these rapid tests are maybe half as good in picking up viruses. Our response to that was we stopped using the rapid test. We paused it. We looked internally at our own data where we had rapid tests and then outsourced labs to look to see, was there really half the rate of positives when you corrected for the test modality—and the answer is no. Then we looked at the study from Langone and it was based on fewer than 30 patients from what I recall. At the time, we looked at our own data set, we had on the order of thousands of patients, perhaps around tens of thousands of patients. So we had a much larger data set. And then we went to Abbott because we were one of their largest customers. And we said, show us the data that you supplied to the FDA to get approval. And their data set was also much larger than Langone. That Langone study came out a few months ago. You haven't seen any repeat studies to say that the rapid test was inferior. 

Basically the reputation of the test got tainted by a very small study that wasn't following the protocol so even the research scientist said, ‘Yeah, that's right, this isn't what the protocol is for the rapid test,’ which is pretty immediately running up the sample once you've collected. So it hasn't been replicated at all and it's not true in our own internal data set either. So we're very confident that it's a good test. We connected our brand to theirs, that's how confident we are. We unpaused it, after we were able to really look at the the data that were out there

Do you give employers an online dashboard where they can monitor the status of employees?

Yes. Health privacy is incredibly important. And frankly employers don't want to be health care providers. When I say, look, we are at the layer between the employer and the employee, we're the health care layer, most of the players are like, thank God. We've had employees ask whether the employer has access to their DNA. So really unusual thoughts out there. But reasonable ones, like, ‘Oh, am I signing away my rights, all my health data to my employer?’ The strict answer is no. We function as your physicians in the same way as if you went into the doctor's office. The caveat is that if you have coronavirus, you've agreed to allow us to share only information about your coronavirus status with your employer. They have a dashboard to see how many employees have completed the daily symptom tracker. How many employees are not allowed to go to work because Carbon has said not to. They don't even necessarily know what the symptoms were, what led us to make that recommendation. And then, who's been tested and tested negative. So essentially the dashboard gives them a really quick overview for planning purposes to know who's going to be in the office and who's not.

They actually just see the names of who tested positive or has been asked to stay home? So they know Kevin Delaney is not going to be in, because Carbon has told him to stay home?

That's right. But they don't know that Kevin Delaney had mild symptoms and it also turns out that he is getting chemotherapy. Or we had a scenario where a gentleman who worked at a biotech company has a husband and he didn't necessarily want his employer to know that his husband tested positive. And just as if you go into your doctor's office, we take the role of your doctor. We wouldn't share that information with anyone in the world without your permission. So we are that layer in between the employer and the employee, and it keeps both sides safe.

I've been reading a bunch about the sewage monitoring for the presence of coronavirus, especially on college campuses. Is that something that any organizations that you're aware of or are looking at as part of the things that they're doing?

No. There are some municipalities that have looked into it. The reason that it's less relevant is that we know that people will shed virus in their stool for months after they're no longer infectious. So similar to an antibody test where you can't really make an individual decision if you knew that 30% of your employees had antibodies or tested positive on the antibody test. You might say, oh, well, that's good. Maybe some of them have some degree of immunity, but it's a very blurry vision of what their actual health status is and the risk.

A lot of people are counting on the vaccine to be able to return to normal life and normal operation of companies. When do you see things being normal again, and is a vaccine the key?

We've been misled, is how I feel about it. When you hear about Moderna and Pfizer and AstraZeneca getting a vaccine to us by election day, or this quarter, 2020—it really creates a false narrative. We know that the technologies and the vaccines that we're trying to rush to market have never, at least the most prominent ones, have been shown to be effective. So to get a vaccine to market in record time on a platform that's never been shown to be effective is a leap of faith. 

Even if you were to accept that, because we're throwing so many resources at it, there's the operational component of it, which is then manufacturing that vaccine to the order of 150 million doses or more, and then distributing the vaccine to people who can administer it, whether it's CVS or Carbon Health or places where the vaccine can be administered. Then you need to have enough people who are vaccinated to really even try to make a difference in the transmission rate, then you add to it that they will likely need a second dose of the vaccine. So now it's not just half the country getting vaccinated, it's half the country getting vaccinated twice. 

The soonest we can get back to normal, assuming that everything is perfect, the fastest development of the vaccine, flawless execution in distribution and production, and then flawless administration. And you're looking at a vaccine that's, let's say, 70, 80% effective, which would be absolute home run. In an article that I wrote about this, I said it would be a fall of 2021 for people to really start to start to go back to normal. In my own mind, now I've changed that to we're looking at 2022, unless the virus mutates away from being so dangerous. 

That is a real outcome that doesn't even rely on the vaccine, which is the virus has been around for thousands of years. It sometimes becomes really dangerous like MERS and SARS and now SARS-CoV-2. But then it can also then mutate away and have the predominant strain be less dangerous. And the timeline for those things might be about the same. 

Carbon has built out these very low-profile popup clinics where today we're doing testing, with Alaska Airlines as an example. But we've really created it so that we can also administer vaccines because the nation today has about half the capacity that it needs to quickly administer vaccines. Vaccine capacity is such where like people get their flu shots from September through like February typically. They trickle in and you don't necessarily need to have all that capacity built up. For SARS-CoV-2 you're going to need to have much more capacity for lots of people that come in, and to do it efficiently. And the only way that we see that happening is with software technology, so you can register ahead of time. We can tell you what the risk of the vaccine is. You sign your papers. You show up at 10 at our pop-up clinic in Seattle. We're like, "Oh, are you Mr. Delaney? And has anything changed since you registered?" You get vaccinated and you walk away. That's our vision.

You're saying things aren't going to return to normal until next fall really optimistically and probably later, but the workplace management school management of exposure and infection gets you to something that's closer to normal before then?

Yes, for sure. I look to a country like Taiwan, where life is pretty normal for most people. There's universal masking and all that stuff, but nearly all businesses are open and people go to work. People can go to restaurants. The rate of mortality, there is six or 10 out of 20 million. We should be 150 people dead if we were as good as Taiwan in deploying contact, tracing isolation, mandatory isolation, which is what they've done. So I think that for a company to say, ‘I'm gonna wait for a vaccine,' that's not a good idea because they will be waiting a long time. And there are ways of opening up businesses today that can get both sides of the aisle happy. 

The narrative of where either they're going to open businesses, or we have to shut down is really a false narrative as well. The point is open businesses in the safe way, through strong leadership that provides important real guidelines to businesses on how to open safely. The data are there. All I can say is that we don't need to wait for some miracles to happen. We just need leaders who will make sure that companies abide by it. And frankly the federal government should be supporting businesses and schools to reopen safely. Does the mom and pop grocery store down the street have the ability to do baseline testing for all their employees? They might once, but if we're doing a cadence of once a month, maybe the government should, help. It's way better than paying for unemployment.