Jump to content

Menu

Florida, positivity rate, reporting, what the heck?


Recommended Posts

Ok, I sort of posted about this before, but what the heck? Can anyone explain this, and is this common other places? Florida Department of health consistently posts a positivity rate that is about HALF what John's Hopkins says! 

I thought maybe John's Hopkins just is reporting in a crazy way, and therefore all states look to have high positivity, but no! Ohio has about 3-4% on the John's Hopkins graphs, but Florida is around 13%! Yet our DOH claims around 5%!

Is there this much discrepancy all over the country? Or do most places use a standard method to calculate this, that matches up with John's Hopkins, and my state is just being disingenuous? (not surprising)

You can see in the graph below Florida actually reports two different rates, calculated differently, but neither is anywhere near the rate John's Hopkins has up. 

 

 

Screen Shot 2020-09-14 at 12.31.37 PM.png

Screen Shot 2020-09-14 at 12.32.07 PM.png

Link to post
Share on other sites
4 minutes ago, Ktgrok said:

Ok, I sort of posted about this before, but what the heck? Can anyone explain this, and is this common other places? Florida Department of health consistently posts a positivity rate that is about HALF what John's Hopkins says! 

I thought maybe John's Hopkins just is reporting in a crazy way, and therefore all states look to have high positivity, but no! Ohio has about 3-4% on the John's Hopkins graphs, but Florida is around 13%! Yet our DOH claims around 5%!

Is there this much discrepancy all over the country? Or do most places use a standard method to calculate this, that matches up with John's Hopkins, and my state is just being disingenuous? (not surprising)

 

I haven't been following it closely, but an Iowa newspaper (the Dubuque Telegraph Herald) has been reporting on the discrepancy between the numbers being reported by the state and the actual positive rate, which is quite a bit higher.

Link to post
Share on other sites
On 9/15/2020 at 2:43 AM, Matryoshka said:

Oh, I'm sure it has nothing to do with these kinds of shenanigans being reported:

[Pic deleted by moderator]

 

This was my first thought!  I can't say more without being completely political!   It is getting harder and harder NOT to be political.   Everything these days is political, including a VIRUS!

 

  • Like 9
Link to post
Share on other sites
1 minute ago, DawnM said:

 

This was my first thought!  I can't say more without being completely political!   It is getting harder and harder NOT to be non-political.   Everything these days is political, including a VIRUS!

 

Well, to be fair I blame a different politician - our governor. These are the DOH reports, not CDC, so..yeah. Blame DeSantis. 

  • Like 1
Link to post
Share on other sites
3 minutes ago, Ktgrok said:

Well, to be fair I blame a different politician - our governor. These are the DOH reports, not CDC, so..yeah. Blame DeSantis. 

LOL, that's why I said 'these kinds of shenanigans' - I think there are others following the leader, iykwim.

  • Like 4
  • Thanks 1
Link to post
Share on other sites
9 minutes ago, Ktgrok said:

Well, to be fair I blame a different politician - our governor. These are the DOH reports, not CDC, so..yeah. Blame DeSantis. 

 

Yes, him too.   I have heard nothing but complaints.   My friend is a teacher in FL and is so angry.

  • Like 1
Link to post
Share on other sites
8 minutes ago, Jean in Newcastle said:

I don't know how you can divorce the two when one follows the other in lock-step. 

Well, only in the sense that he could act like a grown up and choose not to do what his ring leader says. But he doesn't. Like I tell my kids, if you follow along you are guilty yourself. sigh. 

  • Like 1
Link to post
Share on other sites
58 minutes ago, Ktgrok said:

Well, only in the sense that he could act like a grown up and choose not to do what his ring leader says. But he doesn't. Like I tell my kids, if you follow along you are guilty yourself. sigh. 

 

Sure, but in that case the follower and the leader both share the blame - it makes no sense to say "I blame this guy, but not that one, whom he follows willingly off a cliff".

  • Like 1
Link to post
Share on other sites

I have no idea where John Hopkins is getting their data. Where are they collecting it from?

I trust our state data. I mean they could be off, it is a government after all but they have been pretty competent through all of this. I don't know if I would trust your state's data. No offence.

Alaska State DHSS website says 2%. John Hopkins says 3.8%. The highest rate ever according to DHSS was 3.31% in March.

???? I don't know.

Link to post
Share on other sites

I believe JHU counts non-new positives and AB positives in their figures? I have seen data for AZ that shows differences (pasted below), but not FL. I think on the chart you posted FL specifies what specifically they are showing the rate for?

 

20200914_121106.jpg

  • Like 2
Link to post
Share on other sites
1 hour ago, Tanaqui said:

 

Sure, but in that case the follower and the leader both share the blame - it makes no sense to say "I blame this guy, but not that one, whom he follows willingly off a cliff".

Oh, I have all sorts of feelings bout that guy, but he's not directly responsible for my state's data. That's what I meant. The buck stops at the Governor's desk for that. 

Link to post
Share on other sites

I should clarify that it isn't just John's Hopkins that has the higher numbers. Is it everywhere I can find data that has the higher numbers, and they agree with each other. Only my state DOH report is different. 

And yeah, seems my state counts only the first positive, but counts all repeat negatives. So if you test positive once, then have two negatives to go back to work, you end up with a personal positivity rate of less than 0, it seems. Lovely. 

Edited by Ktgrok
  • Confused 3
  • Sad 2
Link to post
Share on other sites

Ok, so John's Hopkins, Covid Action, etc are doing a straight "number of positive tests done that day, divided by number of total tests done that day". Straightforward, can be used the same for all areas. 

Florida has all sorts of mathematical gymnastics involved (oh, and if the person isn't a "resident" they don't count them in a lot of the stats...even though we have a HUGE number of people that live here 5 months a year and who absolutely should be counted if they catch it here...not to mention the tourists, etc)

  • Like 2
  • Confused 1
Link to post
Share on other sites
2 hours ago, Ktgrok said:

Ok, so John's Hopkins, Covid Action, etc are doing a straight "number of positive tests done that day, divided by number of total tests done that day". Straightforward, can be used the same for all areas. 

Florida has all sorts of mathematical gymnastics involved (oh, and if the person isn't a "resident" they don't count them in a lot of the stats...even though we have a HUGE number of people that live here 5 months a year and who absolutely should be counted if they catch it here...not to mention the tourists, etc)

That doesn't seem to be an accurate picture of positivity, though. It seems like what FL is doing is way more specific and transparent and more broken down as to what actually matters for public health (not sure what you mean by gymnastics).

Edited because I was wrong!

Edited by EmseB
  • Like 1
Link to post
Share on other sites
1 minute ago, cintinative said:

The person in FL who set up her own competing dashboard--does she report positivity?  I wonder what her website reports.

Hers lines up with everyone else (but FL DOH). 

Just now, EmseB said:

That doesn't seem to be an accurate picture of positivity, though. It seems like what FL is doing is way more specific and transparent and more broken down as to what actually matters for public health (not sure what you mean by gymnastics).

I didn't fully grasp that JHU is including AB tests in their positivity metrics plus repeated tests until I went looking because of this thread. Wow.

What? I was just at their site and they specifically say they do NOT include antibody tests. 

And the problem with Florida and everywhere else doing it their own way is there is no way to compare from place to place. Also, it is NOT accurate to count positives once no matter how many you get, but negatives more than once. 

  • Like 1
Link to post
Share on other sites
32 minutes ago, Ktgrok said:

Ok, so John's Hopkins, Covid Action, etc are doing a straight "number of positive tests done that day, divided by number of total tests done that day". Straightforward, can be used the same for all areas. 

 

The below is copied from Johns Hopkins (bolding mine) - they are definitely not doing the calculation you describe. They only use negative tests in their calculation.

https://coronavirus.jhu.edu/testing/individual-states/florida

 

Positivity Rates: Our calculation, which is applied consistently across the site and predates most states’ test positivity tracking efforts, looks at number of cases divided by number of negative tests plus number of cases.  ... Because states do not all publish number of positive and number of negative tests per day, we have no choice but to calculate positivity via our approach.

Edited by CinV
to add link
Link to post
Share on other sites

Our governor said a couple of days ago that our positivity rate was about 1%.

J.H. has it listed as 3.17%, or 3x as high. 

ETA Just saying that numbers are off here, too.

Just looked, and the rate is down to 2.87% according to out state health dept. so it’s in line with J.H. I don’t know why the gov. said that it was so low. Maybe it wasn’t the positivity rate, maybe he was referencing the number... what’s it called... that shows the replication rate? I don’t think that that’s the right term. 

 

 

Edited by Fifiruth
Link to post
Share on other sites
18 minutes ago, Fifiruth said:

Our Progressive Democrat governor said a couple of days ago that our positivity rate was about 1%.

J.H. has it listed as 3.17%, or 3x as high.

 

Conspiracy!

  • Like 1
  • Haha 1
Link to post
Share on other sites
47 minutes ago, EmseB said:

Conspiracy!

Yeah right,  🙂

It seems like we are hopelessly suspicious, and thinking that there is a conspiracy on the right to downplay numbers, or on the left to artificially inflate them. 

In this case, I just checked and our rate is down to 2.87% which is more in line with the J.H. number. I don’t know why the gov. said that it was so low. It could be that I misunderstood him. If I can find a clip, I’ll post it an see what you all think.

Link to post
Share on other sites
2 hours ago, Ktgrok said:

Hers lines up with everyone else (but FL DOH). 

What? I was just at their site and they specifically say they do NOT include antibody tests. 

And the problem with Florida and everywhere else doing it their own way is there is no way to compare from place to place. Also, it is NOT accurate to count positives once no matter how many you get, but negatives more than once. 

Ack, I think I was reading the chart wrong! I can edit my post! I was wrong!

However, FL isn't the only state that differs from JHU.

I can't find where it speaks to FL, but this article suggest JHU counts differently than at least a few states, MD and AZ being two. This article about MD says this:
 

Quote

 

Chan said the state chooses to report tests per day because they felt that was most consistent. She also said the state is in communication with Hopkins and others about reported data.

On Thursday, Dr. Jennifer Nuzzo, the lead epidemiologist for the Johns Hopkins Coronavirus Resource Center's Testing Initiative, said Maryland is reporting in a transparent and comprehensive way and the discrepancy in numbers can be, in part, attributed to the fact that Hopkins uses one calculation to accommodate the wide disparities in reporting nationwide and can't customize calculations for each individual state.

 


The article about AZ says that JHU doesn't count some negative tests from some sites but the health department does.

Quote

 

AZDHS said in an interview with Arizona's Family that their numbers are lower because they do not include numbers from testing sites that do not report the number of negative test results while Johns Hopkins does.

A) The AZDHS assumes the results from these facilities have about the same average positivity as the rest of the sample and so discards them while the Johns Hopkins methodology treats these clinics as having a 100% positivity in their testing.

B) The Johns Hopkins reports treat all of the unreported negative tests from these facilities as if they never happened while the AZDHS methodology treats all of the reported positive tests from these testing sites as if they never happened.

C) The AZDHS will rarely over report the positivity rate and can be useful as a lower bound while the Johns Hopkins data will rarely underreport the positivity rate and can be useful as an upper bound of actual positivity.

 

 

Link to post
Share on other sites
8 minutes ago, EmseB said:

Ack, I think I was reading the chart wrong! I can edit my post! I was wrong!

However, FL isn't the only state that differs from JHU.

I can't find where it speaks to FL, but this article suggest JHU counts differently than at least a few states, MD and AZ being two. This article about MD says this:
 


The article about AZ says that JHU doesn't count some negative tests from some sites but the health department does.

 

That's good they clarified!

Link to post
Share on other sites

Wasn’t there an issue with the backlog of tests?  Maybe one set of data back filled it and the other lot dumped it on the day it came in?  Otherwise yeah there’s questions over the morbidity and mortality data being raised and I’ve noticed the decline in deaths doesn’t seem to be following as closely to the decline in cases but that’s probably more to do with the way the virus acts.

Link to post
Share on other sites
13 minutes ago, Ausmumof3 said:

Wasn’t there an issue with the backlog of tests?  Maybe one set of data back filled it and the other lot dumped it on the day it came in?  Otherwise yeah there’s questions over the morbidity and mortality data being raised and I’ve noticed the decline in deaths doesn’t seem to be following as closely to the decline in cases but that’s probably more to do with the way the virus acts.

If you look at Florida the state reported data is consistently less than the John's Hopkins, Covid Action, etc data. Over long periods of time. 

  • Sad 1
Link to post
Share on other sites
Just now, Ktgrok said:

If you look at Florida the state reported data is consistently less than the John's Hopkins, Covid Action, etc data. Over long periods of time. 

Well I think I know which one I’d be making my decisions off sadly.

Link to post
Share on other sites

I don’t know why the difference, but I’d tend to believe the lower number. Not that it’s scientific, but my boys’ school has been open for 6 weeks, sports included, and only just announced a 4th covid case.  Out of 800+ students, plus faculty. My daughters school is similar. The numbers seem really low, and I’m glad we’re open. We’re in the Tampa Bay area.

  • Like 2
Link to post
Share on other sites

I may be way off in my thinking but:

My state reports a positivity rate for "all tests" and another for "unique individuals tests." The "unique tests" positivity rate lines up pretty well with the John Hopkins, but my state is making decisions based on the "all tests" positivity rate.

In looking at the methodology John Hopkins is using, I wonder if they end up with a number more in line with unique individual tests/new cases rather than those rates including all/repeat tests.

My understanding is that most states are using all tests in their figures and that, at this point, we don't do tons of repeat until negative tests so it's not a big deal. At this moment, however, it is a difference of 2.1% between the two in my state as  whole--there is an even greater discrepancy in my county rates. 

  • Like 1
Link to post
Share on other sites
55 minutes ago, sbgrace said:

I may be way off in my thinking but:

My state reports a positivity rate for "all tests" and another for "unique individuals tests." The "unique tests" positivity rate lines up pretty well with the John Hopkins, but my state is making decisions based on the "all tests" positivity rate.

In looking at the methodology John Hopkins is using, I wonder if they end up with a number more in line with unique individual tests/new cases rather than those rates including all/repeat tests.

My understanding is that most states are using all tests in their figures and that, at this point, we don't do tons of repeat until negative tests so it's not a big deal. At this moment, however, it is a difference of 2.1% between the two in my state as  whole--there is an even greater discrepancy in my county rates. 

Florida publishes two different positivity ratings, but neither of them are anywhere near what John's Hopkins publishses. Not even close. 

Link to post
Share on other sites
6 hours ago, Ktgrok said:

I should clarify that it isn't just John's Hopkins that has the higher numbers. Is it everywhere I can find data that has the higher numbers, and they agree with each other. Only my state DOH report is different. 

And yeah, seems my state counts only the first positive, but counts all repeat negatives. So if you test positive once, then have two negatives to go back to work, you end up with a personal positivity rate of less than 0, it seems. Lovely. 

I read through the charts you posted several times to try to understand them.  When I processed what was there, it made total sense to me.  The top is reporting total case positivity, using multiple tests done daily.  For example if you have symptoms, go to a quick-test place, get a positive result, but  later in the day, you're feeling so cruddy that you go to the hospital, where you test, once in the ER, and then on the COVID floor.  You've had three positive that day.   So the positivity rate for the top graph includes multiple positives for each person.  For the bottom graph it only counts the first positive test  (new case test) for one person, not multiples.  And it should exclude repeat tests on additional days for the same person.  (In counting rates of positivity, each day should only include the total of new first time tests that day, to have a useful gauge of how we're changing over time.  Only the new cases, not ones that are in progress.) 

A personal positivity rate of negative is meaningless, and should not engender any emotion.   Repeat tests on the same person should not be counted in any positivity calculations.   I suppose the number of negatives could be counted separately to yield some kind of "how many people are recovered" rate, but even that would be flawed, because we know that people test positive due to viral fragments long after they are well. 

 

 

Link to post
Share on other sites
8 hours ago, Halftime Hope said:

I read through the charts you posted several times to try to understand them.  When I processed what was there, it made total sense to me.  The top is reporting total case positivity, using multiple tests done daily.  For example if you have symptoms, go to a quick-test place, get a positive result, but  later in the day, you're feeling so cruddy that you go to the hospital, where you test, once in the ER, and then on the COVID floor.  You've had three positive that day.   So the positivity rate for the top graph includes multiple positives for each person.  For the bottom graph it only counts the first positive test  (new case test) for one person, not multiples.  And it should exclude repeat tests on additional days for the same person.  (In counting rates of positivity, each day should only include the total of new first time tests that day, to have a useful gauge of how we're changing over time.  Only the new cases, not ones that are in progress.) 

A personal positivity rate of negative is meaningless, and should not engender any emotion.   Repeat tests on the same person should not be counted in any positivity calculations.   I suppose the number of negatives could be counted separately to yield some kind of "how many people are recovered" rate, but even that would be flawed, because we know that people test positive due to viral fragments long after they are well. 

 

 

But they DO count repeat negative tests in some of the models (state has admitted it)

And neither calculation comes close to the rate reported by John's hopkins and others. 

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...