Read on Twitter
Wait. What did I just read?!
I was doing some sleuthing trying to understand why the total NYC stats are so MASSIVE and I discovered:
**NYC is ESTIMATING #COVID19 hospitalizations NOT ON TESTS but on symptoms that match known cases***
First, hats off. NY has been hit hard:
1/
I was doing some sleuthing trying to understand why the total NYC stats are so MASSIVE and I discovered:
**NYC is ESTIMATING #COVID19 hospitalizations NOT ON TESTS but on symptoms that match known cases***
First, hats off. NY has been hit hard:
1/
35% positive rate
50K cases in NYC alone?!
1500 deaths! :(
A few days ago (every day feels like a week now) I sent out the following tweet asking about odd wording on the NYC PDFs which they used before they came out with their data website:
https://www1.nyc.gov/site/doh/covid/covid-19-data.page
2/
50K cases in NYC alone?!
1500 deaths! :(
A few days ago (every day feels like a week now) I sent out the following tweet asking about odd wording on the NYC PDFs which they used before they came out with their data website:
https://www1.nyc.gov/site/doh/covid/covid-19-data.page
2/
The NYC COVID19 data page came out and I had to ask again... what does this "Every Hospitalized Cases" mean?
This is important because NYC cases are OFF THE CHARTS!
They provide a GitHub data repository (which is how I built last night's posts on date discrepancies).
3/
This is important because NYC cases are OFF THE CHARTS!
They provide a GitHub data repository (which is how I built last night's posts on date discrepancies).
3/
They have a pretty good data team over there because all of their README docs at GitHub are there, it's nicely organized... including this this snippet (screenshot and link):
https://github.com/nychealth/coronavirus-data/blob/master/README.md
4/
https://github.com/nychealth/coronavirus-data/blob/master/README.md
4/
"intense resource and time constraints on hospital reporting systems"
No doubt! We've talked about the dangers of a society demanding real-time pandemic medical data in an election-night style format.
But what's that NEXT PART about?
5/
"...we have estimated the number of individuals diagnosed with COVID-19 who have ever been hospitalized by matching the list of key fields from known cases that are reported by laboratories."
Now, word of caution - maybe this estimation is totally normal but essentially...
6/
Now, word of caution - maybe this estimation is totally normal but essentially...
6/
NYC is estimating hospitalizations and cases by ASSUMING that OTHER patients who visited ERs with SIMILAR symptoms in a database should AUTOMATICALLY be counted as #COVID19 patients
Stop me if I'm wrong but that could be ONE of the factors as to WHY NYC has SUCH a big caseload?
Stop me if I'm wrong but that could be ONE of the factors as to WHY NYC has SUCH a big caseload?
Now look - this may be SOP for a crisis like this... and I love a good regression formula... but
1) WAS it a regression formula? Or just a vlookup? Or worse yet... a filter?
2) How far back did they go? What period of time does "Ever Hospitalized Cases" cover?
Lot of questions
1) WAS it a regression formula? Or just a vlookup? Or worse yet... a filter?
2) How far back did they go? What period of time does "Ever Hospitalized Cases" cover?
Lot of questions
