Each section includes information that I gathered mostly from the Robert Koch Institut in Berlin, translated and then shared with public health officials. Ideas and concepts evolved over time and a recommendation could later be contradicted due to effectiveness of distancing and commercial closure, or due to a lack of the ability to create enough tests or recruit enough public health workers. Generally the RKI was not influenced by politics or commercial endevors to more rapidly open Germany, but subtly there might have been. Some data was also gathered from universities in the UK and the University of Washington in the US. CDC and John Hopkins sites are followed by most everyone, so I tried not to dublicate.
Since I was advising the local public health department, the reference is local, but the discussion applies anywhere.
Klamath County has tested 1382 individuals by Monday, Apr 20. The test used is a PCR for the virus. Under ideal circumstances it is extremely accurate. However, it is very dependent on getting a swab all the way to the back of the nose and make the person gag. It is also dependent on the collectors getting the swab into the medium directly and the lab tech must ensure that the swab is processed correctly and all reagents are correct. The results are 33 positive tests, but we might have missed a couple due to the above errors, so there could be 1-2 false negative tests out of the 1382. The way the testing is being done in Klamath County essentially prevents a false positive. It can happen if a couple of samples are commingled in the lab, but the lab techs are very cautious. However, if a person is being tested 14-17 days after an initial positive test, as is being done in some cases, the swab could pick up pieces of RNA that are left over from the virus. Then the test will be positive, even though that piece of RNA is not the whole virus, so the individual cannot spread the disease to anyone else. So with the viral test there are some false positives and some false negatives, but not very many.
Now let us look at the antibody test. This test is more complicated and looks for the antibody in a persons blood. There are 2 different antibody tests being used in Klamath County. One through Quest Labs is made by Abbott and has a sensitivity of 98% and a specificity of 95%. The other by the Mayo Clinic is closer to 99%, but can take 2-3 weeks to get results, so if wide scale testing is done, the Quest test is more likely the one to be used.
So if you do antibody tests 100 people who have recovered from COVID-19 and actually did make adequate antibodies (15-20% of those who recover from COVID-19 do not make antibodies, or at least don’t make enough to give them protection from reinfection for very long). then an ideal test would have 100 positive results. In reality, only 98 will get a positive test result and 2 people will have a “false negative.” Now lets test 100 people that have never been exposed to the virus and definitely do not have any antibodies against SARS-CoV-2, the virus that causes COVID-19. 95 people will get a result back that is negative, but 5 people will have a positive test, even though they do not have the antibodies and are not immune to COVID-19. Is that a problem?
Let’s suppose that 3 times as many people in Klamath County have had exposure to the virus, but never got very sick, or were totally asymptomatic. That would mean there are 120 people out there who have been exposed. Let’s say that 100 of them actually developed the antibodies. So we test 100 of them and get back 98 positive tests, and 2 false negatives.
Now let’s say these 100 individuals are part of a large study to see how many individuals in Klamath County really have been exposed and we are able to test 10,000 people in Klamath County. Ideally, we should have 100 positive tests and 9900 negative tests; but that is not what we will get. We will have the 98 true positives and 2 false negatives. But of the 9900 that should all have been negative, 495 tests will come back positive. That is 496 FALSE POSITIVE tests.
So when we test a large number of individuals and the actual incidence of the disease we are testing for is small, the false positive results can be huge. Statisticians can calculate the false positive rate of the population tested, but it does not apply to any individual test result. So if we do wide antibody testing, can we tell the people being tested the result? “No!” We have 594 positive tests results, but only 98 are true positives and we don’t really know which ones. We can do repeat testing in 2 weeks and narrow the number of false positives, and then do repeat testing again in 2 weeks and get a better idea, especially if we send the repeat tests to the Mayo Clinic. Then in about 2 months, we will know who the real positives are.
In Biostatistics, we use the term, “positive predictive value” of a test tell us how accurate a test is in the field testing for a disease in a population of random individuals. A test may look good in the lab, but a specificity of 95% is not adequate when testing for a disease that is not widespread.
So the Positive Predictive Value in the above situation is:
_________100 (the true positives)________
98 (true positives) + 496 (false positives) = 17%
So our test is only accurate 17% of the time under these conditions.
31 Mar 2020
First case in US 19 Jan when a man and a woman flew to Seattle from S. Korea, both 35, neither knew the other. Both had been in Wuhan. The disease probably reached S. Korea just a little earlier. S. Korea went into full scale attack, but the US did not have a robust public health response for 2 mo (Die Welt 31 Mar 20).
Germany takes patients from France
Germany death rate 0.8%, but many elderly are ill in hospital and it will spread through nursing homes, so the death rate will rise. The testing capacity is very high and they have tested all medical personnel, many minimally ill and all the contacts.
We need to significantly increase the capacity of the public health capacity.
The sooner we trace contacts, the more quickly we can find contagious individuals and isolate them.
90% of Germans surveyed believe in staying home until Easter 12 Apr
Plan is to develop a plan for decreasing restrictions starting 20 Apr, pending analysis or further statistical data
All should wear a mask when in a supermarket in Austria, but in Germany only if you have a cough or nasal congestion.
PPE should be reserved for medical personnel.
We must continue to protect the elderly and chronically ill
Everyone with any sign of any kind of illness must stay home
Pregnant women must be protected from infection
Elderly should not have contact with grandchildren
A website keeps track of the ICU capacity and ED capacity of every hospital. Ambulances will go to the nearest hospital that has capacity
The pandemic is still spreading and will continue to do so for the next several weeks and months
Our testing capacity in Germany is 350,000 tests per week and we cannot increase that capacity. 1 out of every 237 people can be tested each week. They hope to bring rapid tests to local clinics soon and this will significantly increase the testing capacity.
USA therefore needs 5 times that capacity to match Germany: 1.75 M tests per week
S. Korea tests 4000 per million, the US test 5 per million, Germany 4200 tests per million every week
The RKI is asking that all data be collected as there is currently a discrepancy between their data and JHU.
Israeli security is now using all their high-tech spy capabilities to track the virus
Should cell phone data be used in this emergency? S. Korea and Singapore have done so, the EU is still debating the privacy issues.
The Robert Koch Institute questions the accuracy of the JHU tabulations. By gathering data from multiple centers, some cases are being double counted. Since there is a worldwide undercounting of total cases due to limited testing, this may be insignificant, but whereas in Germany the undercount is minimal, this error might be significant. RKI will try to assure that its count is accurate by asking all State Health Ministries to verify that they have a report from every county before submitting their report. Of more concern is that the death count is overstated in many lands, including the US. Today the RKI shows 61,913 documented cases, JHU 71,690; deaths RKI 583, JHU 774.
87% of deaths in Germany are over 70.
The RKI feels that they need in Germany to be able to do at least 100,000 tests per day with same day results, before they can lighten the restrictions. They plan a, “Hammer and Dance” (a term that seems to be used by several epidemiologists) according to Prof Schimidt-Chanosit. Currently, the Hammer is severe restrictions, the Dance will be similar to South Korea, with intensive testing, isolation, quarantine and data modelling with intensive tracking of contacts. The data will be used to adjust the restrictions differently to each county. In Germany the counties are quite small geographically, and tend to be rural, suburban or urban exclusively, so the economies and amount of personal interaction is quite different from one county to another, but usually homogenous in each county.
Currently they are doing 50,000 tests per day, or about 1% of the population each week.
They are doing PCR on nasal swabs, but also in a large subset doing serum antibody testing and follow-up PCR for the virus in 2-3 weeks after a positive PCR.
They are analyzing the antibodies to see if there is similarity to how individuals respond to an infection. This might eventually lead to the production of convalescent serum for treatment of severe cases.
A rapid cassette test will be available next week from Bosch Labs that can be run in all hospitals and most clinics in about an hour.
A point of care test with results in under 15 minutes is being developed, but false negatives with this ELISA tests is requiring further development and maybe very sensitive to specimen collection methods.
The UK is still looking at some limited form of herd immunity, if the virus is found to be widespread in asymptomatic younger patients, but their testing capacity is lagging way behind and they are running out of reagents, so swabs are not getting processed for up to 2 weeks after the sample is obtained.
A worldwide “Manhattan Project” is needed to develop a vaccine, treatments and strategies that can be adopted to different populations with different healthcare availability and populations densities. As Abiy Ahmed, prime minister of Ethiopia, said, “The current strategy of uncoordinated, country specific measures is unsustainable and myopic . We can defeat this invisible vicious adversary-but only with global leadership.”
The success of social distancing will be proven when the number of infected individuals is doubling every 10 days (currently 3.5 days in Germany). Initially, the very large number of negative tests indicated to the RKI that Germany was finding about 90% of the actual cases. Now that they have reached their maximum capacity for testing, more cases are being missed and the flattening of the curve might actually be misleading.
The latest study shows the virus is found in the sputum for 4 days before any symptoms appear. The virus can be detected in stool and urine for 16 days after first symptoms. On average in most countries, people are ill for 7 days before they seek medical attention, therefore have been spreading the virus for 10 days prior to being tested.
4 Apr 2020
Two contradictory studies: one yesterday from Germany, but not from the RKI (their CDC) indicated that the asymptomatic carrier state was huge and combined with the presymptomatic phase being as much as 7 days, would extrapolate that the number of carriers is huge. This is accepted in places like the US as well as most of Europe, where most testing is at capacity and therefore limited to symptomatic, or in the case of Italy and Spain, severely ill (implying that nearly half the COVID-19 deaths are never tested, and the number of fatalities may be nearly double what is reported); but the RKI feels confident that the testing in Germany up until 24 Mar captured 90% of all cases. Therefore they feel the number of asymptomatic carriers is not huge in the general population. It is documented in China, that asymptomatic children can transmit the disease to their parents when confined for days to a small apartment. Analysis in JAMA by Bai et al showed that in following family clusters in Hubei Province, only 1 in 6 asymptomatic children transmitted the disease to another family member, even though all were confined to a small apartment for weeks. Further epidemiologic data is needed, but studies indicated that less than 25% of individuals who test positive for SARS-CoV-2 never develop symptoms, however 50% or more have only mild symptoms and transient mild fever. It should be noted that due to limited testing supplies (reagents) in Hubei Province in late January, the case definition was changed and many were diagnosed by CT scan alone and not PCR testing for the virus. As further data is collected in Europe, especially Germany and Iceland which have very wide scale testing, the true asymptomatic infection rate can be better determined. Initial studies from Iceland indicate that it could be much higher than 50%. Germany on 3 Apr 20 reached its maximum testing capacity of 50,000 tests per day. They plan to have capacity up to 100,000 tests per day next week, but feel they need 10 times that capacity to really track the virus, determine length of shedding, incidence of asymptomatic state and length of presymptomatic state in different ages groups. So the goal is 1M tests per day for a population of 90M.
The pandemic of 1890 probably was also a coronavirus that crossed species and infected humans, then within a year had mutated to a less virulent form. It may be the HKU1 coronavirus that causes the common cold.
The virulence of SARS-CoV-2 is less than MERS or SARS-CoV, but the infectivity seems to be as high as influenza and the common community acquired HCoV’s that cause 60-70% of URI’s. At this time we need extensive RT-PCR testing for the virus itself to determine the spread and serologic antibody testing to determine current immunity and exposure (better helping to determine the asymptomatic infection rate) and to monitor the length of time adequate immunity lasts in different age groups.
R0 ranges from 2.2-6.5 with most studies indicating 3.28, which can be lowered with physical distancing and decreasing contact within the community. An R0 1.0 will stabilize the spread of the virus, but medical costs will still be exorbitant. The study by the economist Scherbina at Brandeis indicates this would still cost the US $1.43B in direct medical costs and $15B if including lost productivity due to deaths of working adults. An R0 of 0.7 through idealized, but still attainable measures would reduce the cost to only $273M in medical costs and $3B with lost productivity. If such measures had been implemented in early February nationwide, the death toll from COVID-19 would be about 450. It’s too late to implement that nationwide, but the restrictions placed by the governor, have significantly reduced the R0 in Oregon and the limited number of cases to date in Klamath County, make it possible to actually keep the local R0 below 1.0, or even 0.5. The next step is to establish a strong mitigation program, so that after 60-90 days, restrictions can be lifted, but all contacts can be tested and isolated, so that the majority of the population can return to work and school, but positive cases can be kept from sharing the virus with others. The establishment of a robust public health system in the next 60 days is possible, but controlling the influx of disease from commuters, travelers and other transient visitors to the county will require extreme vigilance and a program for isolating truckers and others passing through and notifying public health of visitors staying more than just overnight.
7 Apr 2020
Christian Drosten stated 3 criteria that Germany must meet before reducing any of the restrictions on social interaction and physical distancing:
- Everyone must wear a mask in public.
- There needs to be widespread testing of both the virus itself and antibodies with rapid results and follow-up antibody testing on all positive titers one week later to determine IgM and further follow-up to determine immune persistence of IgG. The IgM antibody test would help determine the false positive rate of the PCR test for the virus.
- An App needs to be created and widely disseminated, that would link to the RKI (equivalent to the CDC), that would collect symptoms and temperatures of all citizens and allow rapid testing, and tracing of 72 hours of contacts for further testing.
Some evidence is developing that SARS-CoV-2 has a very limited amount of mutations despite passing through multiple generations. The earliest from China and most recent in Seattle are only 10 nucleotides different. The ones in New York are remarkably similar. This is good for the development of a vaccine and for the potential length one can maintain immunity, but it also means that the virulence and infectivity of the virus will not change much in the next few years.
Some evidence is developing that the virus might have some seasonal variation, so that the restrictions in the US might be able to be lightened after Memorial Day, May 25; but then will rebound in mid to late fall and we will be back where we are today. Will we be ready?
Study being initiated at Stanford to widely test for antibodies. I believe the fingerstick test they are using is for IgG. Unsure of sensitivity or specificity; will have to wait for article. Suspicion that the virus existed in Wuhan much longer than currently reported by China or the WHO team. Much commerce exists between LA and the Bay Area and China, including a daily direct flight from Wuhan. Therefore, the theory is that the virus may have already been widespread in California and a minor herd immunity exits. That is why the fatality rate is so much lower than NYC, although the living conditions are very different.
16 Apr 2020
The problems faced by Hokkaido, Japan: Being an isolated city, and with travel within Japan very restricted, the leaders of this city decided to evaluate the continuation of the shutdown of most commercial activity. The number of cases of COVID-19 in Hokkaido was small, so it was decided that it would be safe to start opening up. Due to limited testing, Hokkaido did not know that many of the residents there had recently become infected and as they went to shops the virus quickly spread. Within 10 days the local hospital was overwhelmed and the death rate sky rocketed.
27 Apr 2020 I recommend that Klamath County and particularly the Southern half do broad testing to see how widespread the virus might be silently lurking in the community which is under pressure from businesses and some County Commissioners to reopen commerce.
Based on recommendations of the Robert Koch Institut (the CDC for Germany, which has done very well and has a 12 step plan for reopening Germany), I recommend six weeks of widespread testing. 1% of the population should be tested every 2 weeks. Test for the virus itself (PCR) not the antibody, which is not accurate enough and does not tell us who is actually spreading the virus. Testing should not just be voluntary or because of symptoms, but larger employers (especially SLMC and the grocery stores) should request all of their employees, even those furloughed to be tested and the Public Health should set up testing sites in front of grocery stores and try to convince as many people as possible to be tested, even if asymptomatic, or had been tested before (but more than 2 weeks ago). If 200 tests were done every week for 6 weeks, then we would know how much the virus has spread in our community and if less than 10% of these tests come back positive, then businesses could safely be opened with some social distancing. If less than 5% of the tests come back positive, these individuals and their contacts could be isolated, traced and tested. The spread could be contained and all businesses including the theaters could be reopened and the 4th of July parade and activities could be held.
This would mean using 1200 tests over 6 weeks, in addition to tests done by the hospital and clinics based on symptoms or exposure to a known positive test. Can we get enough tests to do that?
28 Apr 2020 Headlines on Die Welt indicate Germany has abandoned its plan to do widespread testing of the entire population: 1% of the population every 2 weeks. At this time, only 9% of tests done in Germany are positive and all tests being done are either for healthcare workers, people with at least one of the six WHO symptoms, or known contacts of those who have been exposed to someone with a positive test. The plan had been to test 1 million residents of Germany every 2 weeks. Germany has ramped up its testing capacity to 800,000 per week, so it seemed doable. They have over 90 labs to run the tests, but only one manufacturer of the test kits. The manufacturer will soon run out of reagents needed to make the kits and the basic chemicals come from China and there are no European sources and the supply is limited by worldwide demand.
The plan is now to instead increase the capacity of local county public health technicians to track, test and isolate all positive tests. Continue testing primarily those with symptoms, and then aggressively test all of the contacts of those who test positive.
So my recommendation to Klamath County is to do wide scale testing as long as we can safely use test kits for community testing and not deplete the supply of test kits available needed to test those with symptoms and any of their contacts. Maybe Grant Niskanen, MD can ascertain how many additional test kits could be made available to the county for wide based testing. If it is 1200, then testing could be done for 6 weeks as recommended. If it is only 200, then testing could only be done for 1 week.
It would be hard to say we are out of the woods if we do less than 400 tests over the next 2 weeks. We do not really know at this time what the baseline of infection is. It would be good to test the children of some known positives to see if they are carriers.
If we can do 400 tests over the next 2 weeks of a broad swath of southern Klamath County (be sure to include Chiloquin) and less than 5% are positive, then asking the governor to let Klamath County reduce restrictions on most businesses would be warranted. If over 10% are positive, then we need to keep the restrictions in place and I’m sure OHA would not recommend any change. If the positives are between 5-10%, then how much we try to request for a decrease in restrictions would depend on how well Jennifer Little and her crew feel they can track and trace all of the contacts of positive test results.
If after loosening restrictions during most of May, the number of new cases of COVID-19 goes down, then a request to open theaters and allow gatherings could be considered and submitted to the state.
Ralph Eccles, DO