MS COVID Hospitalizations

[engie]

In the interest of honest reporting here

Date		Hospitalizations by Graph Addition	Hospitalizations by Pie Chart Percentage
1 Apr 2020 Wed		233		360
31 Mar 2020 Tue		218		332
30 Mar 2020 Mon		195		300
29 Mar 2020 Sun		173		268
28 Mar 2020 Sat		152		235
27 Mar 2020 Fri		124		219
26 Mar 2020 Thu		N/A		185
25 Mar 2020 Wed		N/A		150

<tbody>
</tbody>

My previous reporting has been bastardizing the use of both numbers, the same way that the data on the covidtracking project site has also been(they need to fix it and/or show both methods and/or explain which is a true current reflection and use that one exclusively). My apologies. No intention to mislead. Here are the actual numbers per the MSDH website. Why the discrepancy? I don't know. But it seems we are seeing a relatively smooth linear increase in hospitalizations instead of holding flat how I had been reporting.

This is the most important number(s) that bare watching IMO. As long as this doesn't go exponential, we'll reach a point in a week or two here where these additions flatten out to a constant as people are discharged from the hospital(one way or another) at the same rate they are admitted. Probably 10 days at most from that happening.

 

[My Bru]

Good data. Looking at the added hospitalizations each day they do seem to be going down slightly. From 35 to 28 now. Surprised there aren't any spikes. Maybe with all the variables they have manipulated it to show the average.

 

[engie]

Date		Hospitalizations by Graph Addition	Hospitalizations by Pie Chart Percentage
2 Apr 2020 Thurs		262		421
1 Apr 2020 Wed		233		360
31 Mar 2020 Tue		218		332
30 Mar 2020 Mon		195		300
29 Mar 2020 Sun		173		268
28 Mar 2020 Sat		152		235
27 Mar 2020 Fri		124		219
26 Mar 2020 Thu		N/A		185
25 Mar 2020 Wed		N/A		150

<tbody>
</tbody>

Updated with the bad news from today. Hospitalization rate doubled in a single day after being a pretty consistent and slow increase. It's now been identified in TWENTY EIGHT longterm care facilities around the state. The important thing is still not seeing it go exponential. These next 2-3 days are going to be very key.

 

[dawglawz]

Could you ELI5 those two columns?

 

[BogeyGolfer.sixpack]

Date		Hospitalizations by Graph Addition	Hospitalizations by Pie Chart Percentage
2 Apr 2020 Thurs		262		421
1 Apr 2020 Wed		233		360
31 Mar 2020 Tue		218		332
30 Mar 2020 Mon		195		300
29 Mar 2020 Sun		173		268
28 Mar 2020 Sat		152		235
27 Mar 2020 Fri		124		219
26 Mar 2020 Thu		N/A		185
25 Mar 2020 Wed		N/A		150

<tbody>
</tbody>

Updated with the bad news from today. Hospitalization rate doubled in a single day after being a pretty consistent and slow increase. It's now been identified in TWENTY EIGHT longterm care facilities around the state. The important thing is still not seeing it go exponential. These next 2-3 days are going to be very key.

Reading between the lines, I think one reason Reeves ordered the shelter in place was that our state has had several reports of new cases in nursing homes. Prayers for those families and individuals but hopefully those cases can be contained....

 

[00Dawg]

I was also running some related numbers.

So Alabama's new case rate popped yesterday. Only up 10% from the previous high about a week ago, but disappointing. We'll see where today's numbers come in.


Meanwhile, the local media has started a drumbeat of demanding a statewide shelter in place order, and the general explanation as to why seems to be "because everyone else is doing it".
Back in reality, our cases per capita is much lower than every surrounding state.
Numbers are slightly in flux, but at the moment....
Georgia: 517 cases per million
Tennessee: 472 cases per million
Mississippi: 454 cases per million
Alabama: 274 cases per million


Meanwhile, I also took a look at the last round of projections from the Institute for Health Metrics and Evaluation.
For Alabama, today they projected us to be at 6,103 hospitalized cases (range of 3,641 to 7,769) and 1,349 cases in the ICU (range of 777 to 1,750).
Actual total cases at the moment, 1,336.


I don't have Alabama's total number of hospitalized cases, but Mississippi's hospitalization rate is 19.3%. Applying that to Alabama would be 258 hospitalizations. We know about 33% of hospitalized cases require ICU care.
So, 258 hospitalizations and 85 in ICUs versus the low range of 3,641 and 777.

Interestingly, their little projection system's low range peaked on 4/9 (vs 4/17 for the median), at 5,143 hospitalizations. Alabama apparently has about 5,700 available beds now.

 

[horshack.sixpack]

You can tell by the Alabama graph that they had a couple of days of no testing results getting reported. I suspect that pop is really just an influx of results. Also, keep in mind that today's number reported is roughly an indication of how well we were stopping the spread around 3/20ish.

 

[engie]

Could you ELI5 those two columns?

They are covid hospitalizations in MS by two different things put out by MSDH. I can't reconcile why they have two different numbers and don't know which one is correct.

 

[engie]

You can tell by the Alabama graph that they had a couple of days of no testing results getting reported. I suspect that pop is really just an influx of results. Also, keep in mind that today's number reported is roughly an indication of how well we were stopping the spread around 3/20ish.

Agreed. I'm basically living and dying by the date of onset of symptoms graph because it's factoring all of that out...

 

[horshack.sixpack]

Agreed. I'm basically living and dying by the date of onset of symptoms graph because it's factoring all of that out...

I'd like to see a graph that was constantly being refreshed to provide results on the date of testing. The curve would get filled in a little more sporadically, but as complete daily data sets got filled in, you would have more meaningful data. Give me that graph with a vertical line indicating that everything left of the line was complete with no more outstanding tests.

Also, hopefully, some reporting standards can be put in place for states, since that seems to be inconsistent as well...

 

[dawglawz]

I don't even understand what the titles mean. Like "Hospitalizations by Pie Chart Percentage." But the number is 421? That can't be a percentage of a pie graph since it's over 100%, right? I assume that's supposed to be the raw number I guess but then as you say, why two different charts? Then Hospitalizations by Graph Addition? What exactly does that mean? (These are questions not necessarily directed at you).

I haven't seen other states use metrics like this (they may be, I haven't looked).

It bothers me that they're not dumbing this down. More conspiracy-minded types might think they're obfuscating.

 

[engie]

I don't even understand what the titles mean. Like "Hospitalizations by Pie Chart Percentage." But the number is 421? That can't be a percentage of a pie graph since it's over 100%, right? I assume that's supposed to be the raw number I guess but then as you say, why two different charts? Then Hospitalizations by Graph Addition? What exactly does that mean? (These are questions not necessarily directed at you).

I haven't seen other states use metrics like this (they may be, I haven't looked).

It bothers me that they're not dumbing this down. More conspiracy-minded types might think they're obfuscating.

Hospitalizations by graph addition comes from here(this is hyperlinked from MSDH's website). I'm simply adding up across the board to arrive at that number.
View attachment 15752

Hospitalizations from pie chart percentage comes from here(also MSDH). I'm simply taking that percentage hospitalized and multiplying it by the total number of cases to arrive at the pie chart number.
View attachment 15753

And, like I say, I still don't have any idea why these two numbers aren't matching. They should be. It could be that the percentage on the pie chart is cumulative(everyone that has been hospitalized and not just those currently hospitalized) while the raw numbers on the graph are current. It might not be that. I don't know. Going to keep posting them both until I know the answer....

 

[dawglawz]

Gotcha. OK, thanks.

 

[00Dawg]

Sure.

Still, at some point along the curve, if you're comparing like to like, statistic delivery method timing differences smooth themselves out in the results. If you'd have told anyone in Alabama on 3/26 that they'd have a 10% total increase in daily infection rate from 3/26 to 4/2, they'd have taken it.

 

[SirBarksalot]

North Carolina’s hospitalizations jumped 41% over night from 184 to 259. But I’m not sure they have a good grip on counting these, because it was 204 on 4/1 ; 184 on 4/2; 259 on 4/3.
were 20 people discharged on 4/1?

 

[Cooterpoot]

Like I've been saying, this week was expected to be the beginning of the peak ascent. It's been slower than expected. The last two days were expected to be the norm this week. Next 14 days are it and huge.
Nursing homes are getting hammered. Got a dad in one. There's no way to keep it out and no way to stop the spread.

 

[johnson86-1]

Still, at some point along the curve, if you're comparing like to like, statistic delivery method timing differences smooth themselves out in the results. If you'd have told anyone in Alabama on 3/26 that they'd have a 10% total increase in daily infection rate from 3/26 to 4/2, they'd have taken it.

Not sure I know what you're referring to here, but there are two things that are timing related that aren't going to smooth themselves out. THe numer of tests increasng is obviously going to increase the number of positive cases. THere is also the issues that our ability to produce tests has risen faster than our ability to process them, both of those are going to put upward pressure on the results until we are testing and processing results at the level we want.

Neither of those will be smoothed out. Looking at the chart listing cases by date of onset will at least get rid of the upward pressure from lags in getting results. It won't smooth out the increase due to simply more testing.

 

[00Dawg]

Drastic jumps in testing capability would be outside of like to like.

However, in Alabama's favor, Birmingham had a private lab ahead of the pack on day 1, promising same-day testing results. I don't think they've been able to keep up with demand, but between that and Huntsville being a tech center that jumped in to help, the state started ahead of the testing curve for the region, and I don't see indicators that we've greatly expanded local capacity since the initial wave.

Meanwhile, my reference to 3/26 vs 4/1 on 4/3 is directly addressing the lag in testing. Could there still be some in the system? Absolutely. You're also going to get surges from things like the group of first responders in Mobile who were tested all at once, and 30 came back positive. Wait long enough and look at historical data though, and those things even out.

Would I love to look at cases for Alabama by onset date? Sure. More data is great. I haven't seen that listing anywhere yet, though.

 

[00Dawg]

I'm glad you brought this up.

Not discuss the jump, but it leads me back to the Institute, which is being cited by our Executive Branch.
They show that North Carolina should be between 825 and 1,099 hospitalizations today.
So, NC is at ~31.4% of their prediction.

But, what really caught my eye is that their peak prediction for NC is less than 20% that of Alabama's despite double the total population.

Now, NC's curve started earlier than Alabama's, but their moves to address weren't drastically so (NC closed schools 5 days earlier, Bama closed non-essentials two days earlier), yet their curve is better.
View attachment 15754

Is topography to credit for this, better reactions from the populace, or something else?

Regardless, this begins to poke some holes in the data from the Institute. There's nothing here to explain an 80% drop in their predictions from NC to Alabama...other than Alabama hasn't issued a "shelter in place" order. Is it possible they're building a significant calculation change into their projections based on that one action?

 

[SirBarksalot]

I’m not sure what topography would have to do with it...what are you thinking?
Pretty much most restaurants closed their dining rooms when schools closed, same with theaters etc...
Hopefully we keep the hospitalization rate lower than what’s predicted. But I do know they are starting to look for temporary hospital space in places like dormitories at UNCC.

where are you seeing these predictions?

 

[00Dawg]

My mind went...

to the mountainous area of the western parts of North Carolina keeping cross-community contact down, plus the makeup of outer banks doing something similar, but upon further looking, I don't think there's enough population in those areas to explain the difference, even if that was true.

The Institute's info can be found here:
https://covid19.healthdata.org/projections

I believe they plan to update their numbers on Saturday.

 

[SirBarksalot]

You’re correct. Vast majority of the population is in the 3 metro areas I’d say without looking.
thanks for the link

AL- 5000 projected deaths vs 1500-NC

that seems very wonky.

 

[00Dawg]

Alabama's governor....

Has issued a statewide "stay at home" order. However, it's nice window dressing::
[FONT=&quot]The governor’s order means everyone must stay home except when they need to perform “essential activities.” Those include the need to:

[/FONT]

• Obtain necessary supplies
• Obtain or provide necessary services
• Attend religious services
• To take care of others
• To work
• To engage in outdoor activity
• To seek shelter
• To travel as required by law
• To see family members

So, if the numbers from the Institute come out tomorrow with a drastically changed forecast on the basis of a stay at home order, we'll know they're guessing as much as many of us are.

 

[engie]

Today we’re at 284 and 437. Hopefully this means yesterday was a blip on the radar and not the new reality

 

[SirBarksalot]

NC had biggest one day increase with 300+ new cases. 271 hospitalized. I think that’s 231 per million



Afraid ***** about to get real next 10 days.
Saw people playing BB at the park. Cops must have broken it up cause they were gone when I passed by later. Saw a HomeDepot parking lot slammed full.

 

[bruiser.sixpack]

NC had biggest one day increase with 300+ new cases. 271 hospitalized. I think that’s 231 per million



Afraid ***** about to get real next 10 days.
Saw people playing BB at the park. Cops must have broken it up cause they were gone when I passed by later. Saw a HomeDepot parking lot slammed full.

Home Depot is limiting to 100 people in the store at a time. They have personnel stationed at each entrance/exit with a radio.

 

[SirBarksalot]

Has issued a statewide "stay at home" order. However, it's nice window dressing::
The governor’s order means everyone must stay home except when they need to perform “essential activities.” Those include the need to:

• Obtain necessary supplies
• Obtain or provide necessary services
• Attend religious services
• To take care of others
• To work
• To engage in outdoor activity
• To seek shelter
• To travel as required by law
• To see family members

So, if the numbers from the Institute come out tomorrow with a drastically changed forecast on the basis of a stay at home order, we'll know they're guessing as much as many of us are.

hey Double Aught...they must really be crunching some numbers today. Still waiting on this update.

Ed: 1am Its now 4/5...no update. Really interested in this now.