Thread by @chrisvanderveen

THREAD --

So, in an effort to get a better idea of what is going on with the daily release of @CDPHE data on #COVID19 in Colorado...

I spent a lot of time today trying to get a better handle on data being released...

What I found has left me with a bit of a conundrum...

First and foremost...

There are weaknesses that MUST be acknowledged in the release of data...

Deaths, for example, are not reported as to when the ACTUAL death occurred. When CDPHE learns of a death, it reports it...

2/

But there can be a lag there

For example, patient A dies in Mesa County one week ago today.

County health investigates... maybe test result comes back two days later.

County then reports to state...

CDPHE then puts death on its most current list...

But death was 1 week ago

Another scenario...

CDPHE learns of death from county...

But county does not report death date

(this is apparently happening)

(THERE ARE REPORTING ISSUES HERE...)

Which leads me to my next point...

The bar graphs that are included in the reports -- the ones that look like the one below...

include options for both onset and reporting date...

Onset= Date symptoms started
Reported= Date positive was reported to state

(Sorry... here you go)

reported does NOT equal date death was reported...

So, yesterday for example, the state had "1 more death than previous day"

(if you click on the bar on right side of screen from actual report, it shows you that)

This DOES NOT mean one person died yesterday.

And there lies the rub

The state has told us it is committed to telling the public of deaths (and hospitalizations and case # s for that matter).

but just because it reports, for example, 29 new deaths (as it did yesterday) DOES NOT MEAN there have been 29 deaths in last 24 hrs

I fear I'm even confusing myself here...

But the bottom line, folks, is this data is MESSY...

Really messy...

(Not trying to suggest this is anyone's fault BTW)

But the 4pm release points, in many ways, represent the best point-in-time indicators we've got

9?/

I know what you're thinking...

Chris, why not just group deaths by DATE OF DEATH?

I wish we could... I asked for that specifically today...

But CDPHE apparently is not getting those dates consistently (that is a problem IMHO)

10/

the messy data is indeed leading to confusion (many people are barking at us on this issue)

For example, yesterday @COHospitalAssn reported 1228 patients in CO hospitals w/COVID or COVID-like symptoms...

The same day, @CDPHE reported 1,079 patients...

Confused? Yeah.. Me 2

BUT (you knew there had to be a BUT here)

I honestly don't know what else to do here... Seriously...

The data is so messy the state isn't even able to provide us with race of deaths... (again, it's often left off how it's being reported to state)

If there is a better way for us to release this data, I'm all ears...

Honestly...

Date of onset? Maybe.

Date of report? Not lovin that...

We are trying to get more info... We'd love to know recoveries, for example, but that's not being collected.

Have a suggestion?

@ me
DM me
email me: [e-mail]

But, for goodness sake, please stop acting like we are trying to mislead...

We aint...

cc: @GovofCO ..

data reporting is a problem in this state... Your own @CDPHE is clearly acknowledging it...

Counties are NOT reporting the data uniformly...

There needs to be some standard...

Right now your own dept is getting haphazard data...