Cost of getting sick will make you sick.
A new analysis from the Kaiser Family Foundation estimates that the average cost of COVID-19 treatment for someone with employer insurance—and without complications—would be about $9,763. Someone whose treatment has complications may see bills about double that: $20,292. (The researchers came up with those numbers by examining average costs of hospital admissions for people with pneumonia.)
It’s comparison of slope (rates) that’s what is of interest here. The 33.3% line is a straight line, want to be diving below that, not climbing above.
Using population wouldn’t change the slopes, just shift them. log(cases/population) = log(cases) - log(population)
Animated but explained in great detail for all that are interested
Are there insurance plans with Out-of-Pocket maximums that high? I’ve never had a plan with an out of pocket even close to 20k.
It shows total cases, not cases per million. I think that’s a good choice at the moment.
and getting worse. I’ve been watching that.
Hopefully (wan hope, I fear) due to our starting to catch up on testing.
Here’s one that make on feel a little more optimistic. WA state has been dealing with this the longest and looks like they are doing at least as well as SK.
Gov / private sector together might be off to a slow start but once things get rolling things likely will improve fast. At least that’s my hope.
'kin AY!
Link?
From twitter, you can see the source on the bottom right.
Then Singapore has had the first deaths from the COVID-19 pandemic:
LHL predict more cases and more deaths are likely in Singapore before this pandemic is over
China is also experiencing a new wave of infections from returning Chinese students, workers and tourist, mostly from the Americas, Europe and Middle East:
We have all seen the ‘flatten the curve’ model by now. The model shows the capacity to support sick people as a horizontal line. (I’ll call it the capacity curve.) I propose that line shouldn’t be horizontal but instead a complex curve shaped more like a check mark.
I believe the capacity curve is a function of four main inputs: healthcare workers, consumable material, non-consumable material and logistics.
First, healthcare workers. Initially the healthcare workers will have the same immunity as the general population. None. But as they will be constantly exposed to the virus day after day they will get infected far faster than the general population. In the beginning this will cause many to become sick and have to go into isolation. When a part of healthcare workers go into isolation the capacity to treat the sick will decrease.
The good news is the rather small population of healthcare workers will quickly develop community herd immunity. As that happens their illnesses will decrease and their capacity to treat the sick will return to almost full capacity.
Second, consumable material. This is things like PPE (masks, gloves, disinfectants…) and medications. Initially there will be enough consumables. Eventually the consumables will be used and a shortage will develop. This shortage will increase infections among healthcare workers and reduce treatments to patients. This will reduce the capacity curve.
The good news is that production will have been increasing. Eventually production will outpace use and the capacity to treat the sick, for its input, will go back to full.
Third, non-consumable material. This consists of everything from respirators and ventilators to hospital beds. This is a bright spot. We start with a certain number of non-consumables. They do not get consumed or compromised. As things like respirators are newly made they go into the system which increases capacity. As empty hotels are converted into emergency hospitals the number of beds will increase. Over time the capacity to treat the sick only increases.
Finally, logistics. You could also call it experience or efficiency. It’s all those things. It’s a factory figuring out how to make masks faster, or how to deliver those masks to where they need to be sooner. It’s knowing when to best start and stop treatment of a patient to minimize resource use and maximize survivability. Like non-consumables the capacity input only improves.
So we have a capacity curve with four inputs. Two inputs decrease initially - severely - and then return to full or almost full. Two other inputs start at a given level then increase over time. Put that together and you have a check mark shaped capacity curve.
Here’s the scary thing about this disease and it matches what some folks in Italy have told me. Whenever someone says it’s like the flu show them this.
When it goes bad it goes really bad. What kills you is an unbridled immune response called a cytokine storm.
We were lucky in Washington state to have a governor who rang the general alarm early, and called things like they were. Also, the pandemic hit first in the greater Seattle area, which has a well-read population. They read the news, and took action.
It doesn’t hurt that a lot of citizens here are techies. Tech companies quickly had their people work from home. Most people take the emergency seriously and are self-isolating. Those of us who have to work in a facility have divided up the facility into containment blocks, with workers prevented from traveling between then: if one block is compromised, the others remain uninfected.
Late last week young people were using their spare time to congregate in bars. Very quickly the governor shut those down.
This Monday traffic died to maybe 20% of normal (a guess). By Friday it grew to 50%: what that means I haven’t figured out.
Other phenomena:
the value we are concerned with is the rate of spread not number of cases i.e. the derivative of cases as a function of time. The issue with that is that derivatives will amplify the noise in data set. So before doing a derivative a low pass filter should be applied to the data. A simple moving average filter would suffice for a slow rate of change. A moving average filter is a poor choice if the second and third derivatives are particularly large. The point is raw data does not mean very much on its own it should be processed and presented in the correct context.
There’s accurate and there’s guick-and-dirty. The epidemiologists can deal with the niceties, we’re just spitballing for the next few days.
Confirmed cases. No data on testing. - No links but I saw what I thought was a good number, multiple cases by 6 to get actual. Also saw use a multiplier between 10 and 20.
The New York governor was on TV saying they’re doing more tests per capita than South Korea and China right now. That would explain why their numbers are so high. They could be getting an outsized number of test kits?
As of Friday they had about 11,645 positives for 32,427 tests. That’s 36%. Of course that’s likely skewed due to testing suspected cases over non-symptomatic people.
Has anyone seen numbers showing tests performed by state?
