A lot of people have been thinking about the massive wave of Covid-19 to come in the next month in North America. It is sure to be as horrific as it currently is in Italy and Spain in the USA (New York is already there).
But it is not too early to start thinking about the longer term and the decisions we would need to take to start moving back towards 'normal'.
Note: I have read a lot and know how to do data analysis. But I do not have epidemiological or medical expertise so what I say below should just be considered food for thought.
I am not thinking in Trumpian terms here and putting the economy ahead of lives. But there is, I think a risk of over-reaction. It may be that we need to socially distance until September, but we hopefully do not. And I think we need to prepare to push on the accelerator at the earliest time when epidemiologists give the go-ahead to avoid massive economic and social destruction, and excess deaths due to the fight against Covid-19 (rather than due to the disease itself).
I am reading about organizations cancelling activities in the summer and even well into the fall. My thought is DON'T CANCEL TOO SOON unless it involves overseas travel. I have, for example been urging my own faculty at the University of Ottawa not to cancel summer camps. They may eventually need to be cancelled, but we should wait to the absolute latest possible date before making that decision (even if this puts extreme pressure on the teams organizing them). The argument given for cancelling summer activities now is that it takes 'planning' time to ramp up, hire people and train them. This is reasonable. But it would be much better cancel on short notice (despite the pain of that), than to pre-emptively cancel. I have suggested only cancelling summer camps 30 days before they are due to start (so July ones would be cancelled progressively during June if necessary). If we and other organizations can pull off even half a summer of camps, it will allow parents to get back to work and kids to have less stressful lives. And the employment of staff would not be lost to the economy.
I have heard of whole summer theatre seasons being cancelled, even for August (see https://www.osfashland.org/en/tickets-and-calendar/covid-prep.aspx). Yes, it takes months to ramp up for a large production, but all of these artists are going to be potentially destitute as a result. It might have been better to suspend and then pick up after a 70-day hiatus.
For activities involving international travel, such as conferences, it does seem reasonable to make no plans within a 4-to-5 month time horizon, since nobody will want to commit to flying until some certainty returns, too many people have lost money on travel, and insurance won't cover cancellation any more. Besides, the airlines are going bankrupt. This, of course, poses incredible risks for the September semester at Universities that have a lot of foreign students, or even students from other provinces and states. Hopefully epidemiologists will be able to give good advice near the end of May so overseas travel can resume to some extent in September.
The following scientific paper by Glea et al, entitled "Estimated Deaths Attributable to Social Factors in the United States". Discusses death rates in the US in NORMAL times:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3134519/
The paper calculates that in the US the death rate among people with low social support has been calculated at about 33% higher than among people with adequate social support), with about 21% of the US population aged 25-65 suffering from this. That means that about 65000 extra deaths arise from this cause in this age group each year. Another excess 97,000 deaths are caused by excessively low social support among seniors. So the total amounts to about 160,000 excess deaths in normal times. The paper also shows 130,000 excess deaths among people in poverty, compared with what would be the case if poverty could be eliminated. Clearly many people in poverty also have low social support, so we can't add these figures, but it seems fair to say that these social factors account for something like 200,000 excess deaths a year in the US.
If the poor social support rate or poverty rate go up substantially, say by 33% due to massive job destruction and other factors flowing from how we handle the Covid-19 crisis, that perhaps might cause many tens of thousands of excess deaths.
Total Covid-19 deaths in the US are projected to be something between 36,000 and 150,000 according to this site: https://covid19.healthdata.org/projections
I think we have to look at these figures and realize that after the peak passes, we ought to put our feet on the economic accelerator.
The above figures are for the US, but divide by 10 for Canada perhaps (or maybe by a somewhat larger number since we have a better health care system, better political system, and better social safety nets).
So what policy decisions can be made so that the sum total of deaths from both Covid-19 and the fight against it is minimized. We obviously need to keep up strict social distancing until the peak is well past. But then what?
The longer people are not paid and business cannot function, the greater the destruction of the economy. A year of this and I suspect we would be in a deep depression that would take a decade or more to dig out of. Many societal systems would collapse. Millions of people would die due to causes stemming from poverty and other social issues. Even people in countries with socialized medicine and other safety nets would be massively affected. So a year is out of the question, no matter how bad the epidemic. That leaves three possibilities:
Option 1: Get people back to work and kids back in daycare and school at the earliest point that would avoid overwhelming the health care system, perhaps in the last week of May if current peak projections are accurate. This would likely cause more Covid-19 to spread than in the other options, and some people would be extremely angry when they suffer from Covid-19 after this date or have their relatives die. Many people would have lost two months of income, but we might get away just with a severe recession. Government bailouts might be able to save the majority of businesses, so most people would get their jobs back. On the other hand there might be another wave of social distancing needed for a month or so later in the year or in 2021.
Option 2: Wait a bit longer until case numbers dwindle much lower, perhaps near the end of June when active Covid-19 case rates might go back down to where they were in early March in North America. Kids wouldn't be back in school this academic year, but summer camps could be run so people could get back to work, and some forms of low-travel tourism could resume (camping, local attractions) in the summer. There would likely be some outbreaks here and there and a steady flow of Covid-19 cases everywhere, but many businesses may be able to pull themselves back from the brink. But hardship would be very deep for many people. Even with government support, there would be a lot of economic destruction.
Option 3: Be very cautious, trying to completely eliminate nearly all Covid-19 cases, maintaining most forms of social distancing until late in the summer. I think this would be a disaster. A high percentage of companies, large and small would be bankrupt, with impacts lasting many years. We would be in a depression, not a recession. There would be extensive social unrest and mass migrations; many governments would be insolvent too. Real estate prices would crash. Most people's life savings would deeply depleted if not entirely lost.
For each option, the exact dates would depend on epidemiology, and the possibility of being saved by anti-viral drugs is also something we could hope for. In all three cases, social distancing for the elderly and other vulnerable groups would need to continue for longer.
My thinking is this: Governments should plan for Option 1. i.e. release social distancing and shutdown limitations after about 65 days (or when we get back to the level of infections seen in early March in North America) as is happening in China. Hopefully by this time massive-scale testing will be working well so we can track remaining cases, and also maybe start testing to find out who is immune due to having silently had Covid-19. If this happens in late May, that means that people can ramp up for normal economic activity for the summer months (minus international tourists).
Bottom line: Keep summer plans on hold without cancelling (other than foreign travel) at least until early May, when the situation might become clearer.
Tuesday, March 31, 2020
Sunday, March 22, 2020
Maps of Covid-19 cases - March 22, 2020
[Edit: My next post on this subject is here]
On March 17, I published a visualization of Covid-19 cases per million people. We are now 5 days later, so below is an update using the same scale. This map shows the total number of cases accumulated over time, per million. One can now see that the US situation has rapidly worsened, and is similar to Europe (red) having surpassed China.
Australia, Canada and a few South American countries are now in orange, meaning that they are catching up.
Grey cells are those lacking data (e.g. North Korea).
This is not 'new news'. The importance of this visualization is that it shows trends that are less visible in other visualizations. I have been criticized by some people for publishing data that is already well known. The point, however, is that I am not publishing raw data, or even surprising results, but rather making certain aspects of the data more visible.
I have also been criticized for generating visualizations from data that is inevitably inaccurate. In particular, the source data does not account for 'community spread' cases where no tests have been done. Countries that have been able to do more testing of potential cases will appear to have more cases. But this is the best data we have, and I think it is still useful to see these visualizations.
Next I would like to show a visualization of Active cases per million. This is designed to highlight the forward-looking danger to the population and the health care system. Active cases don't include the cases that have been resolved (by people recovering or dying), and suggests the potential for spread. I have used a different colour coding from the above so readers do not think it is the same type of data.
The very dark countries are seemingly out of control with 500 or more cases per million and, presumably, the likelihood of rampant spreading unless there is very strict quarantine and extensive testing to find all new cases. We see this particularly in Iceland (1574), Switzerland (856), Italy (770), Spain (529), and Norway (422) . Again, this is not new data; we know about Italy and Spain from the news, but the situation in the other black or very dark orange countries is not so widely reported (at least in the Canadian media).
Orange cases are at 100 cases per million, and include the US (117). These countries are in real danger.
The fainter the orange the more likely that, just perhaps, the country could keep this under control with decent isolation methods. Canada is at 38.
It is notable that China has now dropped far down on this scale (as has been well reported in the news). It is at 4. Not everybody is confident that the Chinese situation is accurate, but we can only do the best we can with reported data.
Finally, below is a visualization of Active cases per doctor. This shows the current pressure on the medical system. This is very similar to the above, but a few countries have more risk because of fewer doctors per capita.
Very dark countries have one case or more per doctor. They are in a desperate situation.
Bright red countries have 0.1 case per doctor. This includes Canada. Once again, the US is doing worse than Canada, although it has a way to go before catching up to many European Countries.
The pandemic data for these visualizations comes from the Worldometer website. The population data and number of doctors comes from the built-in Excel data sources, as compiled by Microsoft.
I will periodically update these charts so people can see the progression.
Last week I was super-busy working my Vice-Dean role, helping work on policies and processes for the conduct of final assessments to be undertaken by University of Ottawa students who are working remotely now.
On March 17, I published a visualization of Covid-19 cases per million people. We are now 5 days later, so below is an update using the same scale. This map shows the total number of cases accumulated over time, per million. One can now see that the US situation has rapidly worsened, and is similar to Europe (red) having surpassed China.
Australia, Canada and a few South American countries are now in orange, meaning that they are catching up.
Grey cells are those lacking data (e.g. North Korea).
This is not 'new news'. The importance of this visualization is that it shows trends that are less visible in other visualizations. I have been criticized by some people for publishing data that is already well known. The point, however, is that I am not publishing raw data, or even surprising results, but rather making certain aspects of the data more visible.
I have also been criticized for generating visualizations from data that is inevitably inaccurate. In particular, the source data does not account for 'community spread' cases where no tests have been done. Countries that have been able to do more testing of potential cases will appear to have more cases. But this is the best data we have, and I think it is still useful to see these visualizations.
Next I would like to show a visualization of Active cases per million. This is designed to highlight the forward-looking danger to the population and the health care system. Active cases don't include the cases that have been resolved (by people recovering or dying), and suggests the potential for spread. I have used a different colour coding from the above so readers do not think it is the same type of data.
The very dark countries are seemingly out of control with 500 or more cases per million and, presumably, the likelihood of rampant spreading unless there is very strict quarantine and extensive testing to find all new cases. We see this particularly in Iceland (1574), Switzerland (856), Italy (770), Spain (529), and Norway (422) . Again, this is not new data; we know about Italy and Spain from the news, but the situation in the other black or very dark orange countries is not so widely reported (at least in the Canadian media).
Orange cases are at 100 cases per million, and include the US (117). These countries are in real danger.
The fainter the orange the more likely that, just perhaps, the country could keep this under control with decent isolation methods. Canada is at 38.
It is notable that China has now dropped far down on this scale (as has been well reported in the news). It is at 4. Not everybody is confident that the Chinese situation is accurate, but we can only do the best we can with reported data.
Finally, below is a visualization of Active cases per doctor. This shows the current pressure on the medical system. This is very similar to the above, but a few countries have more risk because of fewer doctors per capita.
Very dark countries have one case or more per doctor. They are in a desperate situation.
Bright red countries have 0.1 case per doctor. This includes Canada. Once again, the US is doing worse than Canada, although it has a way to go before catching up to many European Countries.
The pandemic data for these visualizations comes from the Worldometer website. The population data and number of doctors comes from the built-in Excel data sources, as compiled by Microsoft.
I will periodically update these charts so people can see the progression.
Last week I was super-busy working my Vice-Dean role, helping work on policies and processes for the conduct of final assessments to be undertaken by University of Ottawa students who are working remotely now.
Tuesday, March 17, 2020
Map of Covid-19 cases per million - March 17, 2020
The following chart shows the number of cases per million people in the various countries of world as of March 17th. (For followup showing the same visualization method, see my post on March 22)
This shows the number of cases since the first outbreak (not current cases). It clearly shows that the EU is in fact doing much worse than China. It also shows the UK and US are the are the 'big' countries next in line, closely followed by Australia and Canada.
I will do more posts in the days to come showing how this is evolving and also showing current cases.
This map was generated as follows:
1. I took data of current cases from Bing https://www.bing.com/covid by simply cutting and pasting into Excel. Update: It turns out that Bing was getting most of its data from Wikipedia, which was in turn getting most of its data from here: https://www.worldometers.info/coronavirus/#countries. In future I will use the latter (although Worldometers does not give the 'cases per million' in its table for countries with less than a million).
2. I used Excel's Map charting capability to create the above map. Information about how to do that is here: https://support.office.com/en-us/article/format-a-map-chart-2c744937-a1cc-48f7-bc5e-776497343a29. Population of countries is generated in Excel directly from Bing sources.
Note that Diamond Princess data was attributed to Japan.
Below is a table of data for the countries with 10 or more per million.
This shows the number of cases since the first outbreak (not current cases). It clearly shows that the EU is in fact doing much worse than China. It also shows the UK and US are the are the 'big' countries next in line, closely followed by Australia and Canada.
I will do more posts in the days to come showing how this is evolving and also showing current cases.
This map was generated as follows:
1. I took data of current cases from Bing https://www.bing.com/covid by simply cutting and pasting into Excel. Update: It turns out that Bing was getting most of its data from Wikipedia, which was in turn getting most of its data from here: https://www.worldometers.info/coronavirus/#countries. In future I will use the latter (although Worldometers does not give the 'cases per million' in its table for countries with less than a million).
2. I used Excel's Map charting capability to create the above map. Information about how to do that is here: https://support.office.com/en-us/article/format-a-map-chart-2c744937-a1cc-48f7-bc5e-776497343a29. Population of countries is generated in Excel directly from Bing sources.
Note that Diamond Princess data was attributed to Japan.
Below is a table of data for the countries with 10 or more per million.
| Country | Cases per million | Cases | Population |
| San Marino | 3443.11 | 115 | 33,400 |
| Vatican City | 1000.00 | 1 | 1,000 |
| Iceland | 616.16 | 220 | 357,050 |
| Italy | 462.09 | 27980 | 60,551,416 |
| Switzerland | 323.88 | 2742 | 8,466,017 |
| Norway | 271.01 | 1444 | 5,328,212 |
| Spain | 242.06 | 11312 | 46,733,038 |
| Monaco | 232.59 | 9 | 38,695 |
| Luxembourg | 228.05 | 140 | 613,900 |
| Iran | 199.22 | 16169 | 81,162,788 |
| Liechtenstein | 184.59 | 7 | 37,922 |
| Andorra | 183.78 | 14 | 76,177 |
| Denmark | 177.03 | 1025 | 5,789,957 |
| Estonia | 169.83 | 225 | 1,324,820 |
| Qatar | 166.34 | 439 | 2,639,211 |
| South Korea | 161.66 | 8320 | 51,466,201 |
| Bahrain | 158.79 | 237 | 1,492,584 |
| Austria | 151.21 | 1332 | 8,809,212 |
| Slovenia | 132.08 | 273 | 2,066,880 |
| Brunei | 130.63 | 56 | 428,697 |
| Sweden | 115.60 | 1191 | 10,302,984 |
| Belgium | 107.94 | 1243 | 11,515,793 |
| Germany | 103.69 | 8604 | 82,979,100 |
| Netherlands | 99.59 | 1711 | 17,181,084 |
| France | 98.82 | 6633 | 67,118,648 |
| Malta | 81.67 | 38 | 465,292 |
| Finland | 58.43 | 322 | 5,511,303 |
| China | 58.34 | 80881 | 1,386,395,000 |
| Singapore | 47.40 | 266 | 5,612,253 |
| Ireland | 46.33 | 223 | 4,813,608 |
| Portugal | 42.26 | 448 | 10,600,000 |
| Cyprus | 39.00 | 46 | 1,179,551 |
| Czech Republic | 37.18 | 396 | 10,649,800 |
| Israel | 36.44 | 324 | 8,891,800 |
| Greece | 35.97 | 387 | 10,760,421 |
| Kuwait | 31.43 | 130 | 4,136,528 |
| Seychelles | 31.30 | 3 | 95,843 |
| Latvia | 31.09 | 60 | 1,929,900 |
| Maldives | 29.79 | 13 | 436,330 |
| United Kingdom | 29.54 | 1950 | 66,022,273 |
| Armenia | 24.57 | 72 | 2,930,450 |
| Hong Kong | 21.65 | 162 | 7,482,500 |
| Malaysia | 21.28 | 673 | 31,624,264 |
| Macau | 20.88 | 13 | 622,567 |
| Lebanon | 19.67 | 120 | 6,100,075 |
| Albania | 19.14 | 55 | 2,873,457 |
| Australia | 18.34 | 452 | 24,641,662 |
| Panama | 16.84 | 69 | 4,098,587 |
| United States | 16.00 | 5213 | 325,719,178 |
| Croatia | 15.83 | 65 | 4,105,493 |
| Slovakia | 15.44 | 84 | 5,439,892 |
| Republic of North Macedonia | 13.01 | 27 | 2,075,301 |
| Japan | 12.61 | 1574 | 124,776,364 |
| Canada | 12.23 | 449 | 36,708,083 |
| Bulgaria | 11.45 | 81 | 7,075,991 |
| Romania | 11.33 | 222 | 19,586,539 |
| Saint Lucia | 11.18 | 2 | 178,844 |
| Chile | 11.13 | 201 | 18,054,726 |
| United Arab Emirates | 10.43 | 98 | 9,400,145 |
| Palestinian National Authority | 10.20 | 41 | 4019000 |
Subscribe to:
Posts (Atom)