Daily status, Oct. 6: The blogger ate my homework.
Daily Status, October 6
Reminder: any sections that are unchanged since yesterday are grayed out.
I had a problem with today’s update: I was just about finished when it disappeared. I am experimenting with a new (hopefully more robust) workflow.
My Endorsement for president of the US, consistent with the message of this blog: make decisions based on science: Joe Biden 2020.
The COVID-19 data reported by the Virginia Department of health on October 6, 2020 (based on data collected on Octover 5th by 5 PM) shows that the commonwealth added 625 COVID-19 cases, which is one standard deviation below the 3-week average of 845 cases per day. Over the last seven days, Virginia reported 5666 cases, or 810 cases per day on which works out to 9.7 cases/100K people. This is down from 12 cases/100K several weeks ago, but above the 9 cases per 100K from 1 week ago. This suggests the decrease we are seeing for the last few weeks has ended. Currently the weekly case count is statistically flat. The trends over the last three weeks, though, do still include the decrease, with the case load decreasing at 1.2%/per day, 8.5% per week. 59 of the new cases were associated with colleges, mostly from (136) of those cases are from WL/VMI, VT and JMU.
Regionally, all districts showing are flat to negative case growth. The testing numbers now show the percent positive to below the 5% metric over the last week (4.65%) that is often used to indicate sufficient testing (e.g., is is safe to reopen schools). On one other front, over the last week, we are once again below the metric of 10 cases per 100,000 to enter states like NY without quarantine -- we are at 9.7 cases per 100K, down from 12 or so a few weeks ago, but up from 9 one week ago. I can not predict if we will end up on the quarantine list for NY/NJ again, but it is possible.
Outside of a few smaller cities/towns, the state is doing better than it has since the Hampton Roads surge except we are seeing more localized outbreaks; the hot spots are isolated and probably associated with facility outbreaks. Also, looking at the chance of being exposed in a group of 50 (e.g, eating at a restaurant), almost all of the state is blue or clear, meaning less than a 10% chance of someone else in there being COVID-19 positive (still to high for me, but everyone has different standard or risk they are willing to accept).
In the spring, COVID-19 in Virginia was primarily a concern in the DC suburbs. Over a 3-4 weeks, in late may to early June, NOVA recovered, and for about a month, the disease was under control, to the point where the restrictions were eased. Unfortunately, in eastern VA/Hampton Roads, the easing of restrictions results in a surge in cases, which peaked around August 1 (and resulted in stricter restrictions in that area). Since then, the the exception of growth on college campus, the disease has been stable, except for the rural parts of the state, where safeguards are largely ignored. The trends are clearly visible in the figure below. We notice, that, even though there was an increase statewide in the 7-day numbers, NOVA and Hampton roads have not increases, but NW, SW and Central did increase the COVID cases.
With this, we have completed the transition from COVID-19 being an urban disease to being a rural disease. But, looking at the weekly numbers, it appears that the rural areas, whose caseloads were already increasing, surged with the the influx of a carefree demographic (college students). The optimism I felt last week is now somewhat tempered. Largely because of national stupidity.
Regional growth rates are (in fraction per day) continue to to show improvement. Over three weeks, the trends are better than they were last week, so perhaps this one week is statistical anomaly.
NOVA: .979 -- GMU
Hampton Roads/Eastern VA: 0.967 -- W&M, CNU & ODU
SW VA: 0.982 -- VT & Radford
NW VA: 0.9970-- JMU & UVA
The state as a whole is is decreasing with Rt=0.981
The following table shows the number per 100k for each region. Again, NOVA and Eastern VA are doing the best, and the mountainous regions in NW & SW continue having more cases, but are showing improvements as the college cases are getting somewhat under control, but we are now seeing indications that the college students are infecting the general populations.
The following charts are for the 5 regions. in the chart that shows all regions, The individual line charts show the unfiltered data per day, coupled with the trend lines. The trend lines show the different periods of growth. Early in the pandemic, the different parts of Virginia were functioning largely independently, with NOVA mimicking the northern states, and Hampton Roads mimicking the southern states. But, since September 1, the regions have trended with each other. Note that the nose-dive in the chart today is probably an artifact of the 7-day polynomial filtering I use. The filter is called a Savitzky-Golay filter, and basically is a moving window polynomial filter. At the edge (first and last days of the time series), the filter will over compensate for the trend.
One trend not captured in these plot is the impact of a large (3000 person) political rally in Newport News on 9/25. Newport News was averaging about 10 cases a day. Then, 1 week after the rally, they had 20 cases for two days. That suggests that the Trump Rally may have been responsible for 20 extra cases of COVID-19. And then the next day was the super-spreader event at the White House.
In the trends of case, we can either describe it as slow growth since late June, or flat since early august. Both equally describe the data. Since the second week of Sept, the rates have dropped by 3% per day, or nearly 50% over the period (from an average of 280/day to 140/day), though we may be increasing in the last week.
Currently, all jurisdictions are essentially flat to decreasing
Arlington Co.: 0.986
City of Alexandria: 1.005
Prince William Co.: 0.971
Loudoun Co.: 0.954
The number above is Rt: Rt is an exponential time constant, where the number of cases in a time segment is approximately, n=Ao Rt ^ t, where Ao is the number of cases at the start of the segment, Rt is the exponential growth rate, and t is the number of days since the start of the segment. So, if Rt is greater than 1, it is growing exponentially, if it is less than one, it is decreasing each day.
Another way to look at it, todays number are approximately the growth rate times yesterday's numbers. This is the exponential time constant. Fortunately, the time constants are below 1 and our rate of cases is about 6/100,000 per day. Ideally, we would be 0, but 6 is much better than our peak in which was around 30/100000K.
Looking at the trends, since around Sept. 1, there is a clear downward trend in most jurisdictions. The cause of that is unknown, but may relate to the cooler weather. Alternatively, with school being in session, it could be there is more social distancing. The one exception is Arlington, which may be impacted by the decision to stop enforcing social distancing.
What is most noteworthy of the NOVA Zip code map is the the overall lack of contrast. With the exception of near zero population localities, every part is doing about the same. For Fairfax county, north of I66, almost every area was light blue (0.04-0.08 for the last 2 weeks), but that trend has shifted along the Dulles Corridor, including Vienna (22182) and Reston. The cause is unknow.n. There are small scale variations, but the relationships seem to be more related to locations with more shared living arrangements than anything else. Another way to say this is, while social distancing is hard on all of us, for the lower income people, it may be impossible.
Every part of Northern Virginia is down last week compared with the last month, except for Vienna. The good news is, in the last week, almost all of of Northern VA is showing less than 10/100,000 cases (Alexandria City is the exception). Vienna, which had a post-Labor Day mini-peak, has reverted to its prior status as among the lowest caseload per capita in Northern VA. I am not sure why the surge occurred, nor do I know why Vienna reverted to its prior status, but I will point out that about 1/2 of my readers are from Vienna, according to google analytics and perhaps changed behaviors :).
Growth rate (%/day)
During most of the last 2 months, Vienna and McLean tracked each other, as did Reston and Fairfax. Except, for Labor Day to about 1 week ago where Vienna increased its case load. Now, though, it appears Vienna and McLean are again tracking each other.
It is worth noting that almost every region is seeing small upticks in the last few days, almost as if there was a super-spreader event recently.
I am not updating this today as there is a lot of uncertainty in the region thanks to irresponsible behaviors with unknown impacts. That is to say, the situation for the next week may be tenuous.
I have been thinking about how to quantify and present the local risk here in Vienna. My first thought was Vienna is not a island; people that come to Vienna include residents of Reston, McLean, Fairfax, etc. So, I started looking at the region around Vienna: Fairfax County North of 236, and east of the Fairfax Co. Parkway. From that, I defined a series of Zipcodes to monitor: 22030 22031 22102 22180 22182 22181 22124 20191 22066; I am willing to add/adjust this, and am open to suggestions.
In that range, we are running about 5.8 cases per day per 100K. Now, we know that each case is infectious for more than one day, and the primary risk is the pre-symptomatic/asymptomatic people. To capture that, the magic number is probably the cases in the last 10 days. in the last 10 days, there were 55 cases per 100K people in the local region; the region of people that might go to a Vienna restaurant or coffee shop. That means about 0.055% of the people are probably infectious.
So, if you have close interaction with one person, you have a 99.945% chance of not interacting with a COVID positive individual. But, what if you go into a a place with 16 people? Then, randomly, that is a 99.12% chance of not being exposed or less than 1% being exposed. What about 100 people? The risk of one person being positive is 5.4%. What about a large political rally with 1000 people? The risk is about 60% of being exposed randomly.
So, if you are comfortable with a 1% risk of exposure (for most people my age, that would come with a 1/100000 risk of death; for me, it would be more like a 1/2000 risk of death), then take the risk. I am not comfortable with a 1/2000 risk of death, but would be ok with the lower risk.
As expected, there was an increase in cases among Middle Ages (and young adults) delayed from when the JMU students returned home. There are about about 150 cases from the JMU students. We are now back to the baseline number for that age group. The JMU students infected their parents, resulting in an increase of The concern is that the JMU students will infect their families, and we saw an uptick in the middle aged population, which has since recovered; about 50-100 extra people were infected, which was less than I expected.
Note, I will talk about the age distribution in college communities under "college communities"
Overview: Things are improving at most colleges. I have added the Lexington VA colleges (VMI/Washington and Lee) to the discussion. My data mining by zip code requires I group them together, because they share a zip code.
My process combines the VA Department of Health data and what is reported by the colleges. The report is as of 11:00 ET. I need to point out that the VDH cases may include cases not affiliated with the university as I am using geographic surveillance. It is also worth noting that all assume that students feeling ill are going to health service; I have heard anecdotal reports of people not doing that because they did not want to quarantine. The colleges usually update the dashboards after this post, or on Monday/Tuesdays. Except for Radford, all weekly updates are complete at this time. And the numbers are encouraging (except for UVA).
As the process has evolved, it was clear that I needed to show active cases (e.g., last 10 days) in addition to cumulative cases.
Also, I am assuming the colleges are promptly reporting there numbers to the Virginia Department of Health. As it turns out, W&M is reporting the cases approximately 2 days after updating the dashboard. It is probably a result of the phasing of the process.
Note: I have been tuning the following table to improve the estimated number of cases (the order of operations has changed a bit). Also, I am adjusting the pre-student case olad: to estimate the number of cases from students, I subtract out the average number of cases prior to the student arrivals from the reported cases. If the numbers are large (UVA, VT, JMU & Radford), the impact is minimal. This makes it harder, though, to identify cases in more urban settings, like ODU, VCU and CNU.
YELLOW means there may be community spread; still ambiguous
GREEN means no evidence of community spread
BLACK means they went online.
Last 10 days
% that had COVID***
** estimated number of cases is an attempt to normalize for testing limitations. Specifically, I assume at 5% positive, 100% of the cases would be caught. so I normalize it to that value. If the % positive is very high (>40%) I am likely overestimating the numbers.
*** Dashboard cases are only counted if I can find the dashboard. In some cases, it is difficult to distinguish positive tests from cases (1 case may have multiple positive tests; that is mostly at VT). I include active cases if reported, otherwise, I use total cases.
****% population uses the total reported number of students rather than just those on campus; it may be off when the percent positive is above >40%.
***** Old data, not updated for today.
****** Active cases, not total cases
Large Scale Community Spread:
JMU is online for now, but returns to in-person on 10/5. I am worried because their plan does not include testing the entire population now; here is their plan. My concern is the result of the case count at JMU to rise, but slowly. JMU has been fairly transparent with the situation, but could not get ahead of it. At this point, it is mitigation. Harrisonburg is still running 20% positive. They sent the healthy students home (and some sick kids home). The positives, if known, were allowed to stay on campus, which is critical for society as a whole. It is worth noting that sending the on-campus students home seems to correlate with an increase in cases in that age group in Northern VA, and later by the age group of their parents. JMU did not require testing prior to arrival on campus. JMU plans to reintroduce on-campus activities soon -- this time with about 10x the quarantine space and a plan for prevalence testing. Note that over the last two months, JMU students resulted in about 85% Harrisonburg's COVID-19 cases.
We have seen the prevalence of the virus in Blacksburg increase with the return of the students. Each day, there are 30 and 100 new cases validated. There could be more, but some people are just assuming they have COVID 19.Virginia Tech tested only on-campus students. In the last 2 months, VT student cases, as reported in the dashboard, has accounted for 66% of all of Blacksburg cases. But -- VT is only reported the student health center results; as such the cases in 24060 are a better indication of the stats.
With the students return, there has been a marked increase in cases. Week over week, UVA has been increasing by 25% in the case load. In the last two months, 76% of the case load in Charlottesville. I expect the numbers to start to drop, probably by the end of next week, as UVA has imposed much stricter restrictions on students, but so far no impact has been seen.
W&L / VMI There is clearly an outbreak occurring at these schools. The numbers are low because the schools are small. My surveillance approach is based on ZIPCodes, and they are in the same ZIPCode, so they are combined into a single entry. This is just an artifact of how things are tracked, and not a judgement on either school.\
William & Mary -- so far so good -- No evidence of community spread. They now have a reportable number of cases, with the dashboard showing 18 total cases as of yesterday, up from less than 10 7 days ago. That means since Wed there have been six new cases. It is worth noting that W&M is quick to update the dashboard; they update prior to reporting the VDH.
W&M tested all students prior to arrival in town. Williamsburg is a case where the town/tourists are a bigger threat to the college, than the college is to the town. Over the last two months, only 16% of the cases were from W&M students, but the students are 40% of the town population.
Radford was moved to the watch list, as the percent positive has improved. Radford updates the dashboard once a week, which is not sufficient in my opinion, but fortunately they report to VDH regularly. As such, the data on Radford's dashboard is now 7 days old. While it is not how many students at Radford contracted COVID-19. It would not surprise me to see enough that the community, when isolated effectively has herd immunity (meaning over 60-80% contracted COVID-19). The only way to know would be antibody studies. Hopefully, VDH will look into that. But, currently, it seems that Radford has a manageable number of cases. While only 57% of the cases in the last two months are attributed to students, there were so few cases prior to the students return, this means either the dashboard is undercounting, or there were a lot of non-students infected by students.
VCU seems to have the virus under control -- They have beaten back an outbreak, the % positive is good, and there are few new cases. It was promoted from RED to YELLOW. and now Green. The numbers have shown no significant increase in the last several days; quarantine and isolation space is becoming more prevalent. Being in an urban setting the zip-code and regional surveillance that works well at some of the other schools is not particularly helpful here. So, I have to rely on the dashboard.
CNU -- Students have been back upwards 1 month. A few cases. Under control. Note that the VDH numbers for the zip code indicate an increase in cases, but that is possibly the result of other factors, such as a large ill-advised political rally.
ODU -- Nothing noteworthy. My concern with ODU is they did not test the students, so there may be asymptomatic/presymptomatic spreaders on campus (see JMU). Yesterday, they updated the dashboard to show 59 cases on campus, which is about what I would expect based on non-tested students. But the dashboard has not been updated since Sunday. They update it once per week. We will see if they infected other students.
UMW -- Nothing noteworthy. A few cases but they just returned. I am concerned because they did not test all students.
GMU -- Nothing noteworthy. GMU tested all students.
When I started talking about communities, the focus was on the safety for incoming students. Unfortunately, that concept as changed. Now, we are seeing the colleges impacting the surrounding communities. If we look at the age distribution of cases in the communities of JMU, Radford and VT (New River, and Central Shenandoah health districts), we see that, starting about 4 week ago, number of cases for non-college age citizens is increasing -- about two weeks behind the college age curve. This suggests the disease is infecting the general population. So far, it looks like about an extra 265 middle aged and seniors have been infected were infected; we have gone from 12 cases to 22 cases per day in those groups. Furthermore, in the 5 months from the beginning of the epidemic to early Sept, there was an average of 10 deaths per month; in the last three weeks there have been 19 deaths. This is the concern. This trend has been observed elsewhere in the country, where the college students infect the more vulnerable populations.
Since Donald Trump has been infected, or more accurately, with the infection of Trump, we have learned:
2. Even if everyone who interacts with you is tested immediately prior to the interaction, you can still get sick (see 1).
3. If you have an event with hundreds of people, there is potential for it to be a super-spreader event, even if you test everyone at the event.
4. If you do this frequently enough, you will get unlucky.
5. A non-spaced, but outside event is still dangerous.
This has two impacts for us: first, some of the attendees and support people live in our community. I have no information as to the impact this has had. Second, we now know that any crowded situation is a potential spreading event, even if outside. This includes sporting events, concerts, and worship. Please think about this before planning activities.
I have reason to believe about 100 or more of the last week's Virginia cases are a direct result of the combination of the Newport News Trump rally and the Whitehouse outbreak. If we are placed back on the NY quarantine list, it is because of these political activities.
1) You can repost / share in the entirety by forwarding the link, 2) If you want share partial content, you must receive my permission – I need to make sure you understand what I am saying. If anyone sees this work being used without attribution, please let me know as soon as possible. I am willing to have an informed discussion / debate on my approach, but I want to make sure the proper context is captured.
Source data is from the Virginia Dept of Health COVID Site
My Data Approaches
How I monitor colleges from zipcodes.
Why I did this: About the blog
Why I think the death rate is about 1%
I have thoughts on the spread in college communities which can be seen here. This is the same link at the top of the write-up.
Other Sites: John's Hopkins
Kids can pass covid to parents: Pediatric SARS-CoV-2: Clinical Presentation, Infectivity, and Immune Responses
A fun video showing masks work, guy style:
How to wear a mask.
Politics: Some lie.
Donald Trump yesterday said that without blue states, the death rate would be much better. He needs to stop using alternate facts. Red and blue are defined by who the state voted for in 2016. 47% of the deaths come from red states, which is about his percent of the popular vote. But, more importantly, after June first, when states could use science to mitigate the virus, 65% of the deaths are from red states. Facts matter.