UPS truck arrives to deliver one thousand COVID19 antibody test kits to Cloud Medical in Prospect on April 9, 2020
Well I’m glad to report 2 pieces of good news in the midst of a time when most news is grim.
First, as mentioned in previous newsletters, one of the statistics that I am tracking most carefully is the number of COVID19-related hospitalizations. Two weeks ago, the hospitalization rate for COVID19 in Colorado was following an exponential curve (skyrocketing from 80 to 800 hospitalizations in less than 10 days), and although the rate is still going up, it is doing so in a much more linear fashion with 1376 currently hospitalized in our state. While this means that although over 10% of all hospital beds are already occupied by known COVID19 positive patients, which is a lot, there is still a good chance of avoiding the kind of surge that would collapse our in-patient healthcare system.
State health department data shows that “hospitalizations peaked on April 2 with 111 new hospitalizations and there appears to be a significant decline in daily hospitalizations between April 4 and 8”.
That said, most of our local hospitals admit they still do not have the capacity and PPE necessary to perform tests on everyone that they suspect could have the virus, so even these numbers need to be taken with a grain of salt…which leads us to the second piece of good news:
After over 2 weeks and countless conversations with customs officers, the FDA, the president of the biotech company which manufactures the kits which we ordered, and many others, we are happy to report that our 1000 antibody test kits have been delivered to our practice.
We will start our testing program of Cloud Medical members this week and are currently assessing the safest and most efficient protocols for wide-scale testing roll out. So stay tuned for another update coming soon. In the meantime, I wanted to lay out our rationale in greater detail, for those of you interested in better understanding our philosophy, perspective, and intent.
One of the most important factors historically lacking within our healthcare system has been transparency and as a result, trust. It is precisely transparency and trust that Cloud Medical is built upon. So our intent here is to bring you the truth about what we know about COVID19 testing—warts and all. Someone once said that tyranny is the deliberate removal of nuance, and there is a great deal of nuance to be found in rapid viral testing for those who want to dig a bit deeper (read on below).
But at a high level, the antibody test can result in three qualitative outcomes: 1) you most likely haven’t been exposed to the virus, 2) you are likely to be currently infected with the virus and should take extra precautions to self-quarantine (and should confirm with follow up tests), or 3) it’s likely that you have already been exposed to the virus and are no longer actively infected which suggests that you may have immunity to the virus (however, it is essential to confirm with follow up tests in this case as well).
In an ideal scenario, in the face of a pandemic,
we would be conducting wide-scale testing of broad swaths of our population on a regular basis using tests that are 100% accurate, cheap, fast, and easy to administer. This would allow us to track our progress toward the two main desired outcomes over the next 1-2 years as we strive to develop a vaccine which is safe and effective: the dying out of the virus accompanied by a rising tide of “herd immunity” which occurs when lots of people have had the virus, recover from it, and naturally become immune.
But there are a few obvious things that make our situation far from ideal.
- We are completely lacking in population-based testing programs.
- We’re not sure how to scale them, who will conduct them, at what cost, and from where.
- There are questions about which kinds of tests are best to use and when.
- There are also questions about the accuracy of many of the tests (beginning with the original test kits distributed by the CDC to approximately 80,000 patients throughout all 50 states which were found to be “flawed…delivering faulty results”.
At least some of the newer tests kits and the reagent necessary to use them (a great majority of which are being manufactured in China) have also been found to be of such poor quality that some were shipped back to China by angry prime ministers of several European countries, to the embarrassment of the Chinese government.
So, what kind of COVID19 tests are there anyway?
Currently there are 3 main tests available:
- PCR (nasal swab) — which checks for the presence of portions of the COVID19 genome, so since this is a looking for genetic material that is only present in this particular virus, in most cases you can trust a positive result (the test is highly “specific”), but negative results can be wrong in up to 70% of cases (it’s not all that “sensitive”). Our initial wholesale price for this test in February was $275, but the price has now come down to under $100. The average turnaround has been 7 days for a result. Availability is still limited.
- Blood draw Antibody Test — a “serological” (serum based) test which checks for 2 kinds of antibodies in your blood. IgM is the “acute” antibody and becomes detectible within about 5 to 14 days, while the IgG is the “convalescent” antibody and becomes detectible a few days after IgM. The IgG persists much longer than IgM in most cases, but just how long is not yet clear (see below). This test requires a blood draw, costs vary from $25-$150 and has a turnaround of 2-3 days. Availability is limited.
- Fingerprick Antibody Test — also a serological test for IgM and IgG antibodies using a single drop of blood. Cloud Medical has been able to obtain 1000 test kits for a price of approximately $25 per test including supplies. It has a 15-minute turnaround.
Who decided to offer this testing to Cloud Medical members and why?
Originally we asked the question, “how can we do the most amount of good for our community (to limit the surge of hospitalizations and help get people safely back to work) without inadvertently doing things that were in the end futile, or worse, could actually cause harm?” The first answer that came to me was a prototype for a community-based testing program operated neighborhood by neighborhood. I thought that if we could rapidly pull off a testing campaign of all residents in an entire neighborhood (like Prospect for example) and then conduct a feasibility analysis, identify best practices, and ultimately create a blueprint that could then be used by other neighborhoods—well then we’d be on our way to a community-wide and scalable testing program.
I pitched this idea to our small team of volunteers, but after sleeping on it realized that there would be major roadblocks to this idea. How would we get everyone’s name and address and keep everyone’s results private? What kind of HIPAA secure database would be used, and who would run the program? Etc. Quickly it became obvious that there is a natural hub that already houses all of such data and already holds all of those capacities for a large percentage of the population: the primary care clinic.
So we promptly enlisted Cloud Medical as the “idea lab” for conducting a practice-wide testing campaign that would serve as a pilot for other primary care clinics which could then use what we learn to roll out testing on large panels of their own patients and this method, if successful, could be scaled nationwide. We sent out the survey to which most of you responded with overwhelmingly positive results (98.5% voted “Yes”), and placed our order for 1000 test kits.
How were the tests selected?
Initially we planned to use blood draw Antibody testing, similar to the project in Telluride. But this would have required drawing blood in test tubes on a 1000 people, which is very logistically complicated (furthermore, for the purposes of scaling our model, many family practice clinics lack in-house phlebotomy capabilities). We then explored fingerstick antibody tests as a simpler solution and identified UCP Biosciences (located in Silicon Valley with manufacturing operations in Hangzhou, China) which has been developing diagnostic products and technologies for the US market since 2003, already holds numerous FDA (510K) approvals, and manufactures in FDA inspected and ISO 13485 certified facilities. This gave us increased confidence of their existing quality control systems.
Among the many SARS CoV-2 IgM/IgG antibody testing manufacturing operations based in China which have sprung up in recent months and obtained an Emergency Use Exemption (EUE) from the FDA, few seem to have the QA / QC history and rigor in place to the same degree as UCP Biosciences. We were one of the first practices in the US to order through our distributor and tracked the shipment (with all of the hiccups along the way) from Hangzhou to Memphis to Longmont. There were maddening delays and we even enlisted the help of several high-ranking government officials to help cut through the red tape.
What is the “ideal world” scenario for the testing program (used in parallel with social distancing measures)?
In order to limit surges of hospitalized patients and help people know when it is safe to go back to work, our main tool (besides social distancing) is population-based testing. Looking at other nations around the world, Iceland has already tested 10% of its population and half of those who tested positive had no idea they were ever exposed to COVID19 and never experienced any symptoms at all, and Germany has announced its wide-scale testing program after a small town near Düsseldorf recently revealed that 14% of a sample of 500 residents showed positive antibodies to the virus. Given how far behind other countries we are in testing, The Intelligencer published a story today titled,“The Best-Case Scenario for Coronavirus Is That It’s Way More Infectious Than We Think” and pointed out that “until we…institute large-scale serological and ‘community testing’ of that kind, we will be living in darkness”. I am in agreement. We have no idea how many people have been infected in Colorado given there has been virtually no population-based testing, except this preliminary report from Telluride, indicating only 8 of 1631 tests were positive.
There is no doubt that we could be doing more as a country to identify the best manufacturers of diagnostic tests and support them further in scaling up production and distribution. Ideally, we would have public-private partnerships for testing, rapidly deployed supply chains for test kits that have been rigorously evaluated and validated by independent labs, and a thoughtful implementation of “test and trace” methodologies.
At Cloud Medical we see the value of state or county government-run facilities offering testing for those who are not established with a primary care provider, inpatient testing for a large percentage of those admitted to our hospitals or ERs (whether displaying viral symptoms or not), and primary care clinics operating practice-wide testing programs for their patients. Ideally all three settings would be working in coherence, sharing best practices, and communicating openly to provide widespread testing for their local community. There may also be a role for privately funded community-based efforts. In this spirit of collaboration, we have been transparently sharing our plans at Cloud with the Boulder County Health Department and other local stakeholders all along.
It’s clear that there are certain segments of our population who are at particularly high risk for exposure. In addition to medical personnel, we can count police officers, fire fighters and other first responders, grocery clerks, truck drivers, delivery workers, among others as some of the occupations most likely to be exposed. There is merit in targeting testing for both general populations at large as well as high risk groups—understanding the prevalence of the virus in both groups would yield highly valuable information. Cloud has been in contact with fire chiefs and police officers in our area and is reserving a certain number of test kits for them. (Note: many of those who protect our lives and homes from fire and other threats every day are in fact unpaid volunteers. (See below how Cloud Medical is planning to help and how you can join in this effort.)
How much did the tests cost?
The cost to Cloud Medical for the 1000 kits, shipping, lancets, and other necessary supplies came to approximately $25 per test. We are still calculating the exact overhead costs, but it will be very close that number. We plan to charge about $25 per test.
This is a key component of our initiative. We strive not only to prove that it is logistically practical to operate community-based testing by using primary care practices as a hub, we also strive to prove that it is economically viable. Funding always runs out in the end, whether private donations or public grants. At a time when we spend $3.5T per year on healthcare, we must rein in the immense amount of wasteful and unnecessary expense. Especially now. If successful, our project will demonstrate the radical cost-efficacy and fiscal responsibility of office-based testing.
Our goal at Cloud has always been proving a 10X better healthcare experience at half the cost of the cheapest Kaiser plan. It doesn’t take much creativity to charge people a ton of money like most concierge practices do. Our ethos is different. We are champions for your health and your pocketbook.
As a side note: In our opinion we should discourage profiteer resellers of antibody test kits (many of whom are marking up the cost by 5 -10X or more) to people desperate for answers, who may not be as selective about choosing reputable manufacturers of the test kits, and who offer little or no guidance for the various inevitable and complex clinical questions which will arise from testing.
What do we know about the accuracy of the tests?
This is an extremely difficult question and there are as many unknowns as there are knowns. We don’t really have a “gold standard” test other than the PCR nasal swab which, as noted above, is falsely negative in up to a third of cases. But it’s the best we have for now (along with a clinical history, chest X-ray and CT scan findings, etc.) There are several companies who are claiming a 100% sensitivity and 100% specificity on their proprietary tests. I can tell you as a physician that, unfortunately, such things are rare indeed. Most tests, including our antibody kits from UCP Biosciences, are validated against specimens tested by PCR (which brings up immediate questions: what if the specimen being used to validate a test that was negative on PCR but was one of the 30% of the false negatives, and thus is in fact positive?). UCP, in their initial validation trials, compared their test kit with 37 positive and 61 negative specimens and showed an 89.2% sensitivity (which is still better than most PCR tests with lower false negatives) and 100% sensitivity (zero false positives) with their test. We nonetheless interpret these data with caution due to the relatively low number of validating specimens and have personally been in touch with the UCP’s president who has assured us that further validation trials are ongoing, and she will inform us shortly with updated data.
The bottom line is that the SARS CoV-2 virus that causes COVID19 is still so new that there has not been much time for development and validation of diagnostic tools. Our belief is that testing should proceed as swiftly as possible in spite of the various limitations. Indeed, it will be by conducting testing programs that we will be able to sharpen those tools. But we must be very clear and honest about the potential pitfalls of testing. Doing nothing or waiting for a perfect test doesn’t seem like the best option in the midst of the pandemic.
What are the possible results from the UCP antibody test?
There are 2 antibodies, IgM and IgG, so you can either be negative for both, positive for both, or positive for one and negative for the other.
IgM levels begin to rise as early as 5 days and peak about 2 weeks after onset of symptoms but in some studies, less than 40% had any detectable antibodies within 1 week of onset.
By day 15, 100% of the patient sample had detectable antibodies, with IgG lagging behind IgM by several days. Here is a guide which, again, is merely a guideline which may end up changing with newly gathered evidence:
- “Vulnerable” (IgM and IgG Negative): presumed unexposed to the virus and hence vulnerable to getting newly infected—OR—already infected but still in the incubation phase, prior to IgM antibodies developing
- “Contagious” (IgM Positive, IgG Negative): acute antibody is positive, convalescent antibody is negative, hence we presume they are actively infected and can potentially spread the virus
- “Recovered” (IgM Negative, IgG Positive): acute antibody is negative, convalescent antibody is positive, hence we presume they have been exposed, are now immune, and are less likely to pass it on to new unexposed contacts
- “Indeterminate” (IgM Positive, IgG Positive): both acute and convalescent titers are positive, so we cannot rule out the possibility of contagiousness
What are the results that we should wish for in the Cloud Medical testing program?
Paradoxically, a positive antibody result is cause for a sort of celebration. It’s the most favorable outcome. If you are positive for either antibody it means that you’ve both been exposed, and you are one of the lucky ones who did not land in the hospital or end up on a ventilator. It also means that your immune system is working to develop immunity against the virus. While we still don’t know exact details of our “humoral response” (the way in which our body develops antibodies) to SARS CoV-2, whether in some cases it is possible to re-infected, whether IgG is a reliable signal of immunity, and so on, it is clearly better to have been exposed and survived than to not have yet been exposed at all. This is the goal of herd immunity. In Iceland and Telluride only about 1% of the general population had detectable antibodies, while in Gangelt, Germany about 14% of the general population were antibody-positive. Clearly Gangelt is ahead of Telluride in terms of herd immunity (in so far as we can use detectable antibodies as a determinant of immunity).
The fact is that yes, we do actually want large segments of the population to be exposed to the virus in order to develop herd immunity, but we don’t want that to happen all at once so that so many people hit our ER at the same time that the systems gets completely overwhelmed. Obviously knowing where we are along that spectrum would be useful information! And this is one of the main goals of our testing program.
This brings up the point of how population-based testing could potentially do harm. If those testing positive for IgG on the fingerstick are told that they’re “immune” and free to lift social distancing precautions, but in reality their test is falsely positive (or IgG ends up not being as reliable an indicator of durable immunity against SARS CoV-2 as we’d hoped), then clearly we’d be hurting more than helping. For this reason, it is essential to follow up a positive fingerprick test with a blood draw antibody test.
One other consideration will be to identify those at high ongoing risk for exposure. This includes two main criteria— the number of different people that one comes into contact with (besides immediate family) per day or per week? And how many of those are particularly vulnerable individuals? We all know that it is the elderly and those with other chronic diseases that do especially poorly with COVID19. So we want to identify individuals who are around more contacts in general and vulnerable ones specifically. For this we’ve designed a brief questionnaire to accompany your test kit. Ideally any testing is followed by contact “tracing” in order to be helpful in flattening the curve.
So here again are the antibody-based recommendations:
Negative IgM & IgG: You most likely haven’t been exposed to the virus.
Positive IgM: You are likely to be currently infected with the virus and should take extra precautions to self-quarantine (and should confirm with follow up tests).
Positive IgG: It’s likely that you have already been exposed to the virus and are no longer actively infected which suggests that you may have immunity to the virus (however, it is essential to confirm with follow up tests in this case as well).
We have to recognize that, just as with the PCR testing, there will be some false positives and some false negatives in our test groups. We should not make any decisions based on the result of a single test alone. “Use data, but also use wisdom.”
What can I do to help?
Cloud would like to help support our local fire fighters and police officers and we have reserved a number of testing kits for them. We’d like to offer them as a donation and as a small token of gratitude for all they do to keep us safe. If you would be interested in helping, Cloud is offering a “buy one, gift one” program for test kits.
With gratitude to all of you,
and big thanks to Denise Holtz, MA, our Cloud Medical lead at Peloton without whom we would never have obtained the test kits, and special thanks to our wonderful community-based volunteer task force who have contributed time, energy, passion, and extraordinary brain power to this endeavor:
Seth Miller, PhD, Matt Wilder, PhD, Jess Nichol, Josh Zemel, and Richard Furr
From Your Cloud Medical Team