Category: Covid-19

The IHME website predicts that Covid-19 is now peaking or past peak in New York and a couple neighboring states. Also Washington, California and Hawaii. New hospitalizations in those states are expected to decline close to zero by early May. The rest of the US is 2 or 3 weeks behind that schedule. Most states are predicted to peak around May 1, then decline gradually into early June. That assumes social distancing continues through May.

Websites keep appearing with new ways to look at the pandemic. The Unacast site has a Social Distance Scorecard for every state and county, calculated from cell phone location data. The letter grade is based on total travel, non-essential travel, and “encounter density” (how often people are very close to each other). Since we are ahead of the curve here, I’ll cover how it rates a few NY counties, and how it matches with their actual Covid-19 data.

First up is Tompkins County NY, home to Ithaca, Cornell and TurtleSoft. Population about 100K. Encounter density was near the US average in February, but it dropped steeply Mar 13, when all the students went home. Since then we have been very good about testing and isolating. Many people wear masks now, and the sidewalks are mostly empty. 113 confirmed cases as of yesterday, 84 recovered, 4 in the hospital, 0 local deaths.  It’s calm enough that 100 local health workers went by bus to NYC last week, to help out. I think we did it right.

Next is Tioga County NY, where TurtleSoft first began. Mostly rural, with one medium-size village and a few small ones.  Right now it has 19 confirmed infections and 0 deaths, rising slowly. Its encounter density was low to begin with (25% of US average) and dropped from there.

In contrast is Steuben County NY, two to the west of here. Population similar to Tompkins, but 147 cases and 11 deaths so far. Unacast says they are distancing less well, especially for non-essential visits. Of all the rural upstate counties, it’s doing the worst.

Then there is downstate, which is having a very rough time. As of this writing, New York City has 106,763 cases and 7,349 deaths, with another 100K cases and 4K deaths in surrounding counties. It’s worse than most countries. In normal times, the encounter density is about 3000 times the US average in Manhattan, and 500x in Brooklyn/Queens/Bronx. Those numbers are way down now, but even in shutdown it’s still more crowded than the US average from pre-Covid times.

Low population density does not guarantee low infection rates, but it sure seems to help. That may be good news for the 3/4 of the US that expanded during the Automotive Era. Walkable cities with good public transport are fun to visit, but they’re also better at spreading respiratory disease. So far the more car-centric regions have lower infection and death rates. There’s something to be said for isolating inside a big yard, and traveling in a personal sealed metal box.

San Francisco is an exception to the density correlation: it has high density but low rates of Covid-19. Probably because its mayor shut everything down in early March. Seattle also did fairly well, despite their early start. They tested and tracked cases aggressively.

(Edit Apr 15: this link gives a different take on Covid-19 in rural areas).

Stay safe, folks.

Dennis Kolva
Programming Director
TurtleSoft.com

 

New York State shut down non-essential businesses on Mar 23. A bit more than two weeks ago, but it seems like ages.

Construction was included in the list of essential businesses. Some projects continued for a while, but they are all shut down completely now. Getting UI plus $600 a week was probably the carrot to stop working, and the increase in local infections was the stick.

An employee in a local supermarket tested positive for Covid-19 on Saturday.  So far 500 people have gone for testing because of it, but results aren’t back yet. I checked in at the (different) supermarket where I left N95 masks last week: apparently the manager gave them to her family members. So much for making it safer for everyone else in the county.

People are dying in New York City, but it’s less serious upstate. This site has projections for hospitalizations vs hospital beds for each country and state, assuming current social distancing. It predicts that deaths in NYS will peak today and then drop steadily to almost zero at month end. Last week the site was pessimistic about ICU beds for most states, but it’s looking better today. Hopefully their math and assumptions are accurate.

I went back to Cornell 2008 to 2013 and finished a degree in Molecular Biology. It’s fascinating to read the scientific journal articles that cover SARS and Covid-19.  Coronavirus in general has many sneaky ways to bypass our robust anti-viral systems. It won’t be easy to create a vaccine against it.

In a sense, the world was lucky with Covid-19. It definitely has been scary enough that nobody will ignore its family in the future. Good thing, because SARS had a 10% death rate and MERS 34%. This pandemic could have been a whole lot worse.

If you read European history, from the Stone Age until the 20th Century it was a constant stream of war, epidemics and famine. However, the last big war ended in 1945, the last big epidemic was 1909 and the last famine in USA or Europe was 1814. Since then, most of us have had a comfortable, peaceful, prosperous time. We have smartphones, Internet, laser levels, cordless power tools. It’s unsettling to have death sniffing up close again.

But, here we are. It’s time to think about where to go from here.

Dennis Kolva
Programming Director
TurtleSoft.com

 

 

 

Construction estimating uses a lot of math, but it’s simple stuff.  Multiply unit costs times quantities, add them up, and you’ve got a bid.  There might be a little subtraction or division here and there, but nothing more complicated. Construction accounting is even simpler. Thank goodness for sales tax, which at least has percentages to make it more interesting.

As a math nerd, it’s exciting to work with more than just basic arithmetic. So please forgive me as I indulge in a bit more Covid-19 amateur epidemiology.

Recently I mentioned exponential growth, with a series that starts out 1-2-4-8-16-32.  That’s powers of two, which is important for software developers. However, it’s not the only type of exponential curve.

Globally, Covid-19 probably has an R₀ basic reproduction number of about 2.7. With that multiplier, the series goes 1-3-7-20-53-143-387-1046. Seven jumps to 1000, not ten. At 4 or 5 days per jump, it’s roughly 1000x per month. Two months makes a million, where we are at now globally. Three months tops a billion.

That R₀ value is just an average. The actual growth rate depends on the virus, which can mutate. Usually diseases evolve to become more infectious and less lethal, but that’s not guaranteed. It also depends on the human hosts. If they are densely packed and mobile, the actual R is higher. If socially isolated, it’s lower. The whole point of ‘flattening the curve’ is to reduce R down to something close to 1, so the expansion rate is not so scary.

The series for a R₀ value of 1 goes like this: 1-1-1-1-1-1. Still new infections, but they are offset by older cases that either recover or die. Rabies is a disease that acts like that. Nasty, but uncommon. Good thing, because a rabies that spread like Covid-9 would truly be the Zombie Apocalypse, except everyone would end up dead instead of undead.

Unchecked epidemics start out exponential, but eventually they start to run out of hosts. The R value declines to 1, and then all the way to zero.  Overall, the cumulative case count makes an S shape, sometimes called the ‘logistics curve’. It grows steeper and steeper for the first half, then flatter and flatter for the remainder. The number of new cases each day makes a bell curve.

The good news is, the rate of increase globally has started to slow down. Last week I predicted a million cases by Monday, but it took until Thursday instead. Most likely, social distancing is starting to work. Not everywhere, but the data looks promising for much of Europe, and a few parts of the US.

Sadly, much of the US is still in the early growth stage, where it looks easy peasy. In a week or two or three it won’t be.

Dennis Kolva
Programming Director
TurtleSoft.com

 

Last year, I spent many weekends with a heat gun, removing lead paint off the exterior of my house (built in 1910). Construction unit cost = 12 square feet per hour. Bought a box of 3M N95 respirator masks for the work, then misplaced it over the summer and bought another.

Covid-19 has pretty much turned the remaining masks into gold nuggets. I gave a few to high-risk friends.  Offered the rest to my GP, but they are already well-stocked, and using face shields as well as masks. The local hospital needs tons of masks, but folks have mobilized to sew them. A few N95s won’t make much difference there.

There’s good evidence that Covid-19 has a stealth mode: spread by people with no symptoms, via small droplets in the air from talking or breathing. There’s also good evidence that general use of face masks helps to reduce viral spread. China and the Czech Republic saw lower infection rates when masks were used widely.

It’s complicated, because protective gear is in desperately short supply. Where should they go?

Most people in Ithaca are staying home and/or physically isolating, but the grocery stores are still busy. I finally decided to donate the remaining masks to checkout staff at the local Aldi’s supermarket. They are crowded, with no self-checkouts. Maybe masks will slow down the transmission rate, and reduce the number of people who need the hospital.

Construction companies often have face masks or full respirators laying around.  It might be a good time to think about how best to allocate them, based on local needs.

Since masks and other PPE are in short supply, they will need to be reused and sanitized.  Low heat in an oven is the most convenient way to do that. One study says a pre-heated oven at 56° C (133° F) for 90 minutes, 67° C (153° F) for 60 minutesm or 75° C (167° F) for 30 minutes. A second study says 56° C (133° F) for 15 minutes is enough to kill the virus. A third study recommends 60° C (140° F) for 30 minutes. Take your pick.

If you have enough masks, you also can rotate them without additional treatment. The SARS-CoV-2 virus has a half-life of about 1.2 hours in air, .8 hours on copper, 3.5 hours on cardboard, 5.6 hours on stainless steel, and 6.8 hours on plastic. 4 days is the recommended minimum wait before reuse.

Dennis Kolva
Programming Director
TurtleSoft.com

The USA became #1 for cases of Covid-19 today. We lapped both Italy and China in a single day. Globally, there are over a half million cases. It’s doubling about every 4 days, so expect a million next Monday.

Here’s a timeline for Ithaca. It probably is a good predictor for how the pandemic will unfold in other communities. I’ll keep it updated.

Mar 1: 0 cases. WHO warns about pandemic.
Mar 2-6: 0 cases.
Mar 7: 0 cases. NY declares state of emergency (76 cases in NYS).
Mar 8-9: 0 cases.
Mar 10: 0 cases.  Cornell bans events with 50+ people.
Mar 11: 0 cases.  Cornell plans switch to online classes on Mar 28.
Mar 12: 0 cases (15 tested).
Mar 13: 0 cases. Cornell ends classes, sends students home. County state of emergency.
Mar 14: 1 case.
Mar 15: 1 case (46 tested).
Mar 16: 2 cases. All NY restaurants, gyms, malls closed (960 cases NYS).
Mar 17: 3 cases (107 tested).
Mar 18: 6 cases (146 tested).
Mar 19: 6 cases (279 tested).
Mar 20: 11 cases. Non-essential NY businesses closed (7102 cases NYS).
Mar 21: 12 cases (449 tested).
Mar 22: 15 cases (451 tested).
Mar 23: 16 cases (592 tested).
Mar 24: 18 cases (666 tested).
Mar 25: 23 cases (731 tested).
Mar 26: 32 cases (1009 tested).
Mar 27: 48 cases (1191 tested, community spread).
Mar 28: 56 cases (1191 tested, 2 hospitalized).
Mar 29: 70 cases (1197 tested, 2 hospitalized).
Mar 30: 73 cases (1419 tested, 1 hospitalized, 1 recovered).
Mar 31: 76 cases (1419 tested, 1 hospitalized, 28 recovered).
Apr 1:    80 cases (1480 tested, 1 hospitalized, 33 recovered).
Apr 2:    87 cases (1553 tested, 2 hospitalized, 46 recovered).
Apr 3:    93 cases (1620 tested, 2 hospitalized, 51 recovered).
Apr 4:    95 cases (1626 tested, 2 hospitalized, 54 recovered).
Apr 5:    98 cases (1637 tested, 1 hospitalized, 57 recovered).
Apr 6:  102 cases (1988 tested, 3 hospitalized, 66 recovered).
Apr 7:  103 cases (2133 tested, 3 hospitalized, 69 recovered).
Apr 8:  105 cases (2246 tested, 5 hospitalized, 73 recovered).
Apr 9:  107 cases (2274 tested, 5 hospitalized, 76 recovered).
Apr 10: 112 cases (2290 tested, 4 hospitalized, 82 recovered).
Apr 11:  113 cases (2403 tested, 5 hospitalized, 82 recovered).
Apr 12:  113 cases (2408 tested, 4 hospitalized, 83 recovered).
Apr 13:  113 cases (2419 tested, 4 hospitalized, 84 recovered).
Apr 14:  116 cases (2506 tested, 4 hospitalized, 87 recovered).
Apr 15:  118 cases (2546 tested, 4 hospitalized, 90 recovered).
Apr 16:  119 cases (2606 tested, 5 hospitalized, 93 recovered).
Apr 17:  121 cases (2783 tested, 7 hospitalized, 94 recovered).
Apr 18:  121 cases (2808 tested, 7 hospitalized, 94 recovered).
Apr 19:  123 cases (2820 tested, 6 hospitalized, 94 recovered).
Apr 20:  123 cases (2897 tested, 6 hospitalized, 94 recovered).
Apr 21:  123 cases (2947 tested, 6 hospitalized, 94 recovered).
Apr 22:  124 cases (2996 tested, 5 hospitalized, 94 recovered).
Apr 23:  129 cases (3056 tested, 5 hospitalized, 94 recovered).
Apr 24:  129 cases (3095 tested, 6 hospitalized, 94 recovered).
Apr 25:  130 cases (3104 tested, 6 hospitalized, 94 recovered).
Apr 26:  130 cases (3104 tested, 6 hospitalized, 95 recovered).
Apr 27:  130 cases (3212 tested, 6 hospitalized, 97 recovered).
Apr 28:  130 cases (3299 tested, 5 hospitalized, 97 recovered).
Apr 29:  132 cases (3391 tested, 5 hospitalized, 99 recovered).
Apr 30:  132 cases (3578 tested, 5 hospitalized, 100 recovered).

The hospital only has 8 ICU beds, so things may get dire quickly.  We’ll see.  (NOTE Apr 14: not overwhelmed so far, with enough spare capacity to treat a few Covid-19 patients from NYC).

I’ve been watching Coronavirus since it first hit the news in December, and started to prepare in February. And yet, despite the long advance warning, I probably still should have acted faster. Most likely, today’s 32 cases were first shedding virus around Mar 14. At that time I was still shopping, going to the gym, eating in restaurants. Slightly cautious, but more would have been better.

I have a hand-wavy theory that the human brain is hard-wired to calculate jumps between trees, from when we were small arboreal primates. The distance you fall from gravity is squared over time. The series goes 1-4-9-16-25-36-49-64-81-100. The same math also applies to throwing things. Also cars, and most other stuff we deal with in real life.

Epidemics and explosions are exponential. The series goes 1-2-4-8-16-32-64-128-256-512-1028.  It starts out slower than the gravity curve. Humans often are fooled by that, and think it’s no big deal. Then it catches up. Suddenly it’s 10x bigger. Another 10 jumps and it’s a million, when we were expecting 400. Another 10 jumps is a billion.

Dennis Kolva
Programming Director
TurtleSoft.com

In the last post I did math to figure if and when Covid-19 overwhelms hospitals. The next few weeks will determine how accurate it was. Massive social distancing may prevent a crisis.

There is a second problem that may be more serious, with an even bigger impact on construction businesses (and everyone else).

By now, you’ve probably seen the “flatten the curve” graphic. If the R₀  reproduction value is high, the number of cases is a steep bell-shaped curve. If it’s lower, then the curve is more shallow and the peak is lower. If people aren’t close together, R₀ drops.

With an infection growth rate of 15% and no social distancing, half the US population will test positive for Covid-19 approximately May 31. At 25% growth it’s May 3. At 30% it’s April 26. The US is not ready for that size of widespread disaster. Just look at Italy or Iran. It’s worth the chaos now, to flatten the curve and have the next couple months be less insane.

Problem is, those curves don’t have a time scale.

To avoid high death rates, R₀  has to be low enough to not overwhelm hospitals, and stay that way until nobody is catching the disease. Containment is unlikely with such a highly-infective virus, so it ends when almost everyone has caught it (resulting in “herd immunity”).

It’s not hard to calculate how long it takes to handle Covid-19 safely. It goes like this:

1. Start with US population (327.2 million).
2. Estimate how many will eventually get Covid-19.
3. Estimate how many of those get pneumonia.
4. Multiply that by average duration in critical care (about 14 days).
5. The result is ICU-days. Divide that by the number of ICU beds available (65,000 to 95,000). The final result is days required to flatten the curve.

You’ll notice there aren’t numbers for steps 2 and 3. This is a new disease, so there isn’t good data for those yet.  However, it’s possible to plug guesses into the spreadsheet, and see what results.

Assuming 70% of people eventually get Covid-19 (the best guess I’ve seen online) and 5% of them end up in intensive care (probably close to results from Wuhan and Italy), the curve needs to be spread out over 4.6 to 6.8 years to avoid zapping the health system. If 1% need ICU care, it requires 338 to 493 days. Of course, many folks are working on vaccines, but those will still take 12 to 18 months to test and produce. That’s a very long spell of social isolating.

Hospitals in the US are run for profit, or by non-profits that are chronically underfunded. To cut expenses, they eliminated excess beds and staff that would have come in handy, right about now. Japan has 13 hospital beds per 1,000 people. S Korea has 12, Germany 8, China 4, US 2.77.

I’m really not sure what the consequences will be for small construction companies, small software companies, or anyone.

Dennis Kolva
Programming Director
TurtleSoft.com

 

 

I just made a spreadsheet with New York coronavirus data, to get an idea of how bad it is likely to get here, and when.

Looking at other countries, the most important factor is whether the hospital system is overwhelmed. If so (as in Wuhan, Iran and Italy)  the death rate is high, and it’s a national emergency. If not (as in Taiwan, Singapore, S Korea and most of China), then the pandemic is less of a disaster.

It takes 3 numbers to calculate the risk and timing:

1. 1. 1. Virus growth rate. World-wide outside China, cases are increasing by about 15% per day. NY cases are growing by 30% per day. Calculating from R₀ rate of spread (1.4 to 3.9 infections per case) and incubation period (5 days, range 2 to 14) gives numbers in about the same range.
      2. Hospital capacity. Different sources say 3,000 to 4,000 ICU beds in the state.
      3. Percent of confirmed cases that need intensive care. Most studies show it between 5% and 10%.

The most optimistic calculation says that NY will run out of ICU beds on April 15 (four weeks from today). The most pessimistic version says crisis hits March 31 (two weeks from today).

Different sources say the US has 45,000 to 95,000 ICU beds. Using the same logic, they all are filled sometime between April 3 and April 25.  Four weeks away, plus/minus 10 days. The math assumes that ICU beds are currently half full, but another 50% can be added on short notice.

Keep in mind that testing in the US has been woefully sparse. All confirmed counts are just a bare minimum. Actual infections could easily be 10x or 100x higher. However, that does not affect the timing by much. The numbers it changes just cancel out.

For local planning, many state health departments are now publishing case counts in each city or county. Wikipedia also is adding tracking pages for hard-hit states (search coronavirus + your state name).  Their basic pandemic page is at this link.  (update Mar 19: Politico has tracking of all Covid-19 test results for each state).

There’s a good visual of the NY growth curve on Wikipedia. Take your state’s current case count, and you can guess at what it will look like over the next week or two. All US states now have at least 1 reported case.

Also keep in mind that there’s a week or two lag between infection and testing/confirmation. It means that social distancing does not act quickly to reduce hospitalizations. You need to wait a week or two to see the benefits.

NY/NJ/CT closed all restaurants and bars yesterday. Everyone is panic-buying and preparing to hunker down. This is one situation where irrational fears are actually useful.

According to the math, we may just squeak by. To make it happen, social distancing has to reduce R₀ to 1 or below, and start really soon.

Dennis Kolva
Programming Director
TurtleSoft.com

Not much programming progress this week: our staff has been distracted by pandemic news and planning. It’s hard to stay focused on code when something possibly-deadly is happening outside.

Coronavirus probably has been around ever since vertebrates first moved onto land. It lives in the respiratory system and spreads mostly via cough/sneeze droplets. Several strains exist in humans. They and rhinovirus cause most common colds. No big deal.

However, COVID-19 is a new variation. It jumped from bats to another mammal to humans last fall. People aren’t used to it, and that’s why it’s dangerous. This new virus is more infectious than influenza, and more lethal. Most people just catch a chest cold with slight fever, so they barely notice it (and spread it to others). 20% get pneumonia and need hospital care. Some of them die: especially if over 60, or in already poor health. With good medical care, the death rate is under 1% (still 10x as bad as the flu). When the disease spreads too quickly, hospitals are overwhelmed and 3% to 6% die.

Countries with significant deaths from SARS in 2003 were well-prepared for SARS-CoV-2. They contained the spread, and are almost back to normal now. It required a huge response in China, and less extreme actions elsewhere.

The rest of the world hasn’t done as well.  Things are real bad in Italy right now, and they turned deadly very quickly. The rest of Europe is only a week or so behind them. Conditions in the USA are more uncertain, since there has been so little testing here.  However, cases are already reported in almost every state. Celebrities are catching it. The shit is almost certain to hit the fan very soon. Exponential growth gets scary at this part of the curve.

It’s too late to stop Covid-19, but it can be slowed. Then hospitals aren’t overloaded, and fewer people die. Word is already getting out about the need for social distancing, but here are some thoughts more specific to construction companies:

• • • Shut down for a while, if you can. If you are solo, this is a great time to get some  work done on your own house. If you have employees, can they handle a break? Isolation is the best way to protect your own health, and the health of your community.
    • If not, use face masks. N95 masks significantly reduce the spread of virus-laden droplets. “Easy breathe” masks won’t protect others if the user is infected, but a bit of duct tape over the outlet will fix that. Full respirators and haz-mats are even better if you can stand them. Fortunately, construction sites are less disease-transmittable than most workplaces, and they have other hazards that require masks.
    • Be clear with employees and subs. Send them home ASAP if sick, or don’t even let them come in. This might be time to start paying sick leave, if you don’t already. Emphasize the need for testing and self-quarantine. If you have an ear thermometer, it might help to check people for fever.
    • Stagger work hours & locations if you can. The less time people spend close to each other, the slower disease spreads.
    • Be clear with clients. If you work in client homes, let them know your plans. You might need to add plastic sheeting barriers, and sanitize more thoroughly at day’s end. Switch to outdoor work, if possible.
    • Take the usual precautions. Wash hands often. Cancel group events. Practice social distancing. Listen to local & state health departments. Check out the links below.

What you should do as a Business Owner
Covid-19 FAQs from Harvard U
CDC Covid-19 Website
Vox guide to Coronavirus
Johns Hopkins tracker (sometimes overloaded)

Please stay safe and healthy!

Dennis Kolva
Programming Director
TurtleSoft.com