20 Comments - Ok - Monday, Apr 6, 20 @ 10:30 am: There is a weird thread in conservative twitter right now that says the University of Washington’s model isn’t matching with reality, proving that states like Illinois and New York are exaggerating their need. But their chart “proving” that the reality is less than the model for some reason (and I can’t quite put my finger on it) only shows data from 41 staes and excludes data from Illinois, New York, Washington, and the other hot spots. For some reason. - Grant, Pershing and Eisenhower - Monday, Apr 6, 20 @ 10:47 am: This war will be won the old fashioned American way. Throwing lots of material resources and personnel at the problem. Healthcare resources need to expand for the foreseeable future. The Laissez-Faire attitude of the Trump administration is painful to watch. It’s very clear that they don’t know how Americans fight and win wars. - Generic Drone - Monday, Apr 6, 20 @ 10:51 am: @ Grant… I agree. Can’t wait until Trump moves on. - muon - Monday, Apr 6, 20 @ 10:53 am: I noticed the same jump in peak deaths and looked at the data they had entered since their last update on April 1. On April 1 the IMHE total number of deaths matched the IDPH count at 141. They matched on April 3 at 210. But the IHME model for April 4 shows 64 deaths and a total of 274 deaths. That’s the same number that IDPH has for April 4 and 5 combined. It looks like the IHME model entered data for the two days together. If that was so, it’s likely that the IHME model would predict a sharper rise and more deaths than if the data was spread over two days as shown by IDPH. I would further note that the April 1 IHME model projections for April 4 and 5 were pretty close to the actual counts by IDPH and well within their projected range. Their current April 5 projection of 340 total deaths in a range of 279 to 456 puts the IDPH count outside the 95% confidence range. It really looks like there was a data entry error. - Generation X - Monday, Apr 6, 20 @ 10:54 am: ===But their chart “proving” that the reality is less than the model for some reason (and I can’t quite put my finger on it) only shows data from 41 staes and excludes data from Illinois, New York, Washington, and the other hot spots.=== I don’t know what thread you are referring to, but some of that missing data is likely due to states not reporting hospitalizations. That is a key indicator on assessing both needs and deaths. If you are looking for a critic of the U of W work, Alex Berenson former NY Times reporter has done a lot of work on this and thinks IMHE has way overshot the number for COVID 19 - Anotheretiree - Monday, Apr 6, 20 @ 10:59 am: April 16th feels right. China shows us it takes a 70 day lockdown so the 16th is about midpoint. Agree we need for Trump to “move on” or as I like to put it….”Free us from the scourge of Jared Kushner” - Anon - Monday, Apr 6, 20 @ 11:07 am: Because smoking is shown to increase the consequences of the virus some countries are now prohibiting the sale of tobacco products. It would be interesting to overlay the maps of Covid deaths with tobacco use to see if this could help “bend the curve” - Candy Dogood - Monday, Apr 6, 20 @ 11:19 am: 1.) Look at the range of their certainty. Yikes. Their certainty for the peak day goes up to 8,000 for a range. 2.) How do they get to the underlying assumption that there will be — zero — deaths per day from this starting in June? I understand the benefit of creating a short term model to project needed resources, but I wouldn’t trust this model to make a forecast beyond where we might be in a few weeks, especially if it is relying on assumptions that get us to a point where we have ZERO community spread in the United States. We still have people going to church in states where they have still encouraged those gatherings, and with almost certainty there will be churches violating government orders to meet on Easter. Optimism is okay, but this doesn’t seem realistic. - Candy Dogood - Monday, Apr 6, 20 @ 11:24 am: ===China shows us it takes a 70 day lockdown so the 16th is about midpoint=== Mid point for 70 days starting right before the last week of March? I know it feels like a long time but we’re only st 21 day since the schools closed, unless you want to toss in a weekend day. And Illinois is ahead in time frame for other actions. I hate to be the person advocating this, but this model and this data feels too optimistic. I hope I am wrong, lord do I ever hope I am wrong, but lets prepare ourselves for the high point of their range for certainty. - Moe Berg - Monday, Apr 6, 20 @ 11:31 am: muon: good catch; I hope you will email and let them know Candy Dogood: good point about church in other states; one wonders if by the time it’s all said and done whether Red states, especially Southern ones, are going to come out of this proportionally worse off than Northern Blue ones. - Corn Desert - Monday, Apr 6, 20 @ 12:23 pm: **Red states, especially Southern ones, are going to come out of this proportionally worse off than Northern Blue ones.** Possible, but not likely. Population density plays as much a part in the spread as any other factor. - Moe Berg - Monday, Apr 6, 20 @ 12:35 pm: Corn Desert: not trying to pick a fight, I’m not an epidemiologist or a bio-statistician, however, some people whose job titles are closer to that description just put a piece in the Washington Post: “The covid-19 crisis is going to get much worse when it hits rural areas” - which I saw after my original comment. Excerpts: “Over the past few weeks, our urban centers have scrambled to mobilize in response to the mounting covid-19 cases. But be forewarned: It’s only a matter of time before the virus attacks small, often forgotten towns and rural counties. And that’s where this disease will hit hardest. “Rural communities could fare far worse than their urban and suburban counterparts. Rural populations are older on average, with more than 20 percent above the age of 65. Rural populations also tend to have poorer overall health, suffering from higher rates of chronic illnesses such as heart disease, diabetes and lung conditions, all of which put them at greater risk of becoming severely ill — or even dying — should they become infected. “Rural areas also already suffer from a rural mortality penalty, with a disparity in mortality rates between urban and rural areas that has been climbing since the 1980s. Chronic financial strain and the erosion of opportunity have contributed to ‘deaths of despair’ as well as a rise in conditions such as heart disease, Type 2 diabetes and stroke. Add in prolonged social distancing and the economic downturn, and these trends will surely worsen.” - Chicago Cynic - Monday, Apr 6, 20 @ 12:36 pm: Muon is exactly right. God-willing we won’t be seeing deaths at that rate today. If that’s the case, given the weekend data, they really may need to revise this downwards. - Generation X - Monday, Apr 6, 20 @ 12:41 pm: ===Muon is exactly right. God-willing we won’t be seeing deaths at that rate today. If that’s the case, given the weekend data, they really may need to revise this downwards.=== Returns today in New York are promising. Hospitalizations are down for 3rd straight day and deaths are flat. - Still anon - Monday, Apr 6, 20 @ 1:01 pm: - Ok - The dispute over IMHE’s model focuses on New York. The revised model forecasts that New York would have 24,700 hospitalizations and 6,033 ICU beds in use as of today. The actual numbers are 16,837 and 4,504. In both instances, the actual numbers are just within the very large range of uncertainty. The prior iteration of the model (from April 1) was off by more. - Still anon - Monday, Apr 6, 20 @ 1:03 pm: The concern is that public policy is being driven by models that may be highly uncertain, at best, or highly inaccurate. IMHE is to be congratulated for making its model public. I wish all models being used by federal and state authorities were public, along with the assumptions they used (IMHE does not publish that information). - Chicago Cynic - Monday, Apr 6, 20 @ 3:16 pm: 33 is not good but it’s a far cry from the 82 projected by UW for today. Frankly, it’s good news but it also cries out for UW to revise their model. - Levois J - Monday, Apr 6, 20 @ 4:29 pm: Now I really want to study statistics and data collection. How does any organization project whether it be for pandemics or epidemics or for expected sales? - Candy Dogood - Monday, Apr 6, 20 @ 5:02 pm: ===God-willing=== We’re all in trouble when it comes down to Inshallahs all around. - Candy Dogood - Monday, Apr 6, 20 @ 5:15 pm: === Population density plays as much a part in the spread as any other factor. === I think this kind of attitude implies a fundamental misunderstanding of the current way of rural life in the United States. While population density is definitely a factor, I think failing to understand the extent to which an American living in a “rural” region travels is significant. We have a country that is full of people who commute dozens of miles to work, travel to the nearest Wal-Mart, or stop at the same gas stations, etc. Population density is a factor, but when we travel and gather regardless of where our actual residences are is an issue. Sorry, comments for this post are now closed.