Keeping my head out of
the  Covid 19 sand

As a completely new virus, Covid 19 spread rapidly across
the world, the first and only way to protect ourselves was
to self - isolate in order to stall the spread. By mid April,
t
he data showed that  most countries had succeeded in
“flattening” or “bending the curve”, but with varying
success afterwards.


The first cases in the US occurred on the West Coast,
who were prepared for an infection from Asia and were
able to contain the outbreak.  Unfortunately, New York
was not focused on a possible infection from Europe and
they became the epicenter of the US infection, where it
rose rapidly and threatened the health care system in a
few key areas. Covid  spread to the rest of the US very
rapidly ( 40 states over 15 days), but with large
differences in case count (50x).

By mapping the severity of the infection, and the date of
first death to neighboring states.  It is possible to create a
plausible path for the infection.  Texas was one of the last
states to get infected, probably via Colorado, and one of
the smallest infections in the first wave.


As the infection got established, the general population
started to self-isolate, which was encouraged by a number
of proclamation by state governments. States with later
starts to infection, isolated relatively early, and  has much
smaller infection rates.


The  infectiousness of a virus, is measured by the number
of new cases created by a single infectious individual  
(Infections per person or Ipp). At the start of the infection,
Covid has an Ipp of 5, which is similar to Measles or
Mumps.

In order to follow the spread of the infection,
I created a
modified model of infection which accounted for the
effectiveness of isolation, and the time it  takes for
changes in behavior to impact infections.

The result is a plausible time line for the infection in TX.
Patient Zero for TX was probably asymptomatic and
arrived in the first week in March. Based on changes in
mobility measured by cell phone activity, the TX
population were starting to modify behavior in the same
week. By Mid March, the State had issued “stay at home”
orders. In late March, the change in behavior was
reflected in infection growth. By mid April, the infection
had stalled and started to fall. Because the infection in TX
was relatively small, it appears that people has already
started to relax their self-isolation. In early April,in TX  
infections started to rise again with 1.2 infections per
person.  




It is illustrative to compare the progression of cases in TX  
to NY. In NY, because they were infected  first, the
infection was 30x higher than TX  before isolation had any
effect.


It is clear that self -isolation has been very effective in
managing the infection, at an obvious cost to the
economy.  

There are simple isolation guidelines that can be derived.
At the end of June, in round numbers, we are at 200
cases per day in 1 million people in Travis. Asymptomatic
= 10x symptomatic so 2000 total. Infectious for 10 days,
so 20,000 out of 1 million are infectious. Therefore,1 in 50
people are probably infectious.

If you sit in a room with 50 people, the odds are someone
is infectious. Based on choirs and restaurant examples, if
you share air for 1 hr, there is a real chance you will get
the virus.

No way to guarantee safety except on top of a mountain
on your own. Meet 1 or 2 people for a minute or two with a
mask, the odds are you will be safe.

Beware of bars, rallies, conferences, indoor restaurants,
and of course churches.


How dangerous is Covid ?

The other key property of a new virus is how dangerous it
is, i.e. what fraction of  infected individuals die.   In the
early stages of an infection, a popular measure is the
ratio of Deaths to Cases called the  Case Fatality Rate
(CFR).  CFR values have ranged from 0.5% in Iceland to
14% in Italy with a world  average of 7%.  In the US, the
variation is from 1% in Nebraska,  to 10%  in Michigan,
with a US average of  4%.  In general, these variations
have been blamed on non-uniformities in  detection,
classification and reporting.

The worst fatality rates happened in states with the worst
infections, leading to the suspicion that stress on the local
heath care system has an effect on fatality. Analysis  
shows that keeping  the maximum of daily new deaths
below 10 per million, appears critical. This corresponds to
300 a day for Texas, we are at around 30 today. In NY
and the Lombardy region of Italy, the infection was so bad
that the health care systems started  to collapse.

Michigan also showed unusually high fatalities which
seems to correlate to the state of urban decay in Detroit,  
which was the hot spot of their infection.

The CFR in TX, NY and MI can be broken down as shown
to the right.

The graph emphasizes the importance of controlling the
maximum death rate, and focusing resources on hot spots
in poor communities.

The final Infection Fatality Rate is probably around 0.4-
1.0% due to the large population of asymptomatic cases.

Just like the flu ...NOT !

Here is a  comparison of  the 2019 flu cycle and the  
Covid infection in the US. The model of Covid used an Ipp
= 5.6 ( Ref 4 )  infections per person with no isolation, so
it  followed the  natural disease progression (standard
SIR). Initially,  the Covid infection grew much more  
rapidly  than flu, fortunately, isolation stalled the Covid
infection. The model shows what might have happened
without isolation, the death rate could have risen to a
disastrous level of more than 10,000 a day. Even with
isolation,  Covid still had a death rate 10 x worse than the
flu.









The second wave is here

As a few of the powers that be are at last getting excited
about the "spikes", here is an update that shows that the
recent increases are just a continuation of the exponential
growth since the start of June. There is no new special
event to blame.

This is a log plot, so exponential growth shows up as a
straight line. You can follow the phases of the pandemic
through the linear sections. In March, we had the very
rapid "growth" phase, which forced the lock-downs that
"stalled" the infections in April. In May the infections
"stabilized", and states (NY) with sever infections had hard
lock-downs that let the infection fall. States with mild
infection (TX, FL), had softer lock-downs that allowed slow
rise in infections. In mid-May, the re-openings started, and
in early June the infection started to grow more rapidly
again. Now in late June, CA, KS and MN have all joined
the increasing trend.

In July many states were forced to back off re-opening,
and by the end of July re-stall had started.

The black line shows the level at which most states run
out of hospital beds, Florida, Arizona and some counties
in Texas were in real; trouble by the end of JUly.

.

Understanding the second wave

The only encouragement is that the fatality rates are
dropping, In TX, on June 20,
CFR is 0.7%, meaning that
almost all the infections are occurring in the under 60

community
.  The at-risk population is being careful and
not getting much less infections.



Herd or not to herd

The critical question is now how does this understanding
impact our actions moving forward. There are really only  
2 choices; keep the infection under control and grow the
economy, or get to herd immunity as quickly as possible
without cratering the health care system. Likely range for
herd immunity at 80% is 600K to 1.5M dead nationwide in
0.5-1.5 years.

Choice 1, is maximizing the economy whilst preventing the
infection growing. The emphasis is on testing and
quarantine. As there is a large asymptomatic population,
this means frequent comprehensive testing for the virus
and thorough contact tracing. This maximizes the number
of people who are around to benefit from future
treatments or vaccine. In Texas, we have had increasing
infections for the last month, so Choice 1 is not a option.

Choice 2,
The graph shows the model prediction if NOTHING
CHANGES, or if we re-lockdown around June 1st. The do
nothing scenario takes us to cases and deaths that are
about 2x the worst case NY experience. It does get us
close to herd immunity, but with a total for the US 0.8-1.5
M dead – “The darkest days are still ahead of us.”

We are walking a tight rope to manage the infection. To
maintain an Ipp of 1.6 still  requires the  community to live
their life with less then 50% of the  contacts that we had  
pre-Covid.
Links to more nerdly details
The Data
The Model
The Infection time line
Fatality of Covid
Covid compared to others
To herd or not to herd
The second wave is here
Modelling the second wave  
Managing Covid
Vaccine progress
Therapies
The daily death count for a countries shows who stalled  the infection, and how
effective their isolation strategies have been.
A plausible path for the spread of Covid based on the date of first infection (color
code) and severity of infection (pillar height)  in neighboring states.
The time line of growth in daily Asymptomatic and Symptomatic case count in Texas,
along with isolation measured by cell phone mobility. Texans started to isolate soon
after patient zero arrived. It took 16 days for isolation to impact infection growth.
The time line of growth in daily case count in Texas, compared to NY, showing that isolating
early radically reduces the  infection. Texas is showing the effect of reopening
The Case Fatality Rate increases with Health Care Stress, and the poverty level in
any  infection hot spot. .
A  daily case and death count scenario for TX at end of May, assuming the current level of
isolation is maintained.
The flu and Covid compared. Covid grew very rapidly until isolation stalled the infection.  The
model with no isolation, shows that Covid could have grown to over 10,000 infections a day if
allowed to run its course.