Thread by @juliette_io

So many numbers swirling around this weekend about NSW hospitalisation rates

3 big questions:

1. What on earth is the hospitalisation rate in NSW?

2. How is it calculated?

3. Should Hospital In The Home (HITH) be included?

Let's go into the weeds🌱

#covid19aus #covid19nsw

Q1 - What's the hospitalisation rate in NSW?

The NSW Surveillance Report now tells us this in black and white (it just added this new piece of info last week)

In the most recent report (data to Aug 21), it says it's 11%

Two things to know about this ...

i) Think of it as 'at least 11%' b/c it doesn't account for the lag between new cases and hospitalisations. The report itself acknowledges this

ii) Dr Jeremy McAnulty confirmed yday this rate *DOES NOT* include HITH (sorry for the caps, but that's important)

Q2 – How is the hospitalisation rate calculated?

There are options

(Hate maths? Skip the next 4 tweets)

Surveillance Report calculation =
Total hospitalisations since June 16 / Total cases since June 16

According to The Saturday Paper (TSP), advice to national cabinet calculation =
Ppl needing hospital / Active cases from last week 🧐

This calculation accounts for the lag, so that's good

But we have to assume a couple of things
1. It's referring to people in hospital today
2. It's not using the grossly inflated active cases figure reported by NSW (which takes in 28-30 days of cases) but something more sensible, like a 14-day sum

So if we apply that ...

= Today’s hospitalisations / 14 day sum from 7-11 days ago
= 1030 / 9816 - 11,309
= 9.1% - 10.6%

On Aug 21 (when the surveillance report said 11%)
= 12.3% - 12.8%

TSP reported this calc = 16% but it didn't say when. This calc yields that 3-4 weeks ago

So, it looks like we're in the right neighbourhood

And remember, we're still just talking about hospital in-patients. We haven’t even started on HITH!

So that brings us to Q3 - Should Hospital In The Home (HITH) be included in the hospitalisation rate?

The Saturday Paper reported the hospitalisation rate cited by the Premier, 5.5%, was one-third the real figure because of 'a decision' to exclude HITH

If that's true, that would mean the hospitalisation rate reported by NSW would be very different from other states, right?

What hospitalisation rate does Victoria report?

In 2020 ...
Total Covid hospitalisations in VIC = 2,492
Total cases = 20,368
= 12.2%

(Sources: covid19data.com.au / abc.net.au/news/2021-06-2…)

In VIC today ...
Today’s hospitalisations / 14 day sum from 7-11 days ago
= 1030 / 9816 - 11,309
= 11.4% - 14.6%

What would happen to the NSW hospitalisation rate if we did include HITH?

Today, there would be at least 3156 hospitalisations = 34%

Even if national cabinet advisors were using the ridiculously huge denominator of NSW active cases, the rate would be 24%

For the visually-inclined, here's a chart with no percentages, just raw numbers ...

NSW cases = black line
NSW hospitalisations = green dotted line

VIC cases = blue line
VIC hospitalisations = pink dotted line

They look pretty similar, right?

Note, NSW cases haven't peaked yet, so hospitalisations are still catching up

Now let's add HITH (yellow dotted line)

That is crazy high. Way bigger than anything VIC reports

I thought TSP piece raised excellent points about ambulance services

And the clinical questions of whether some HITH patients should be admitted to hospital are way above my pay grade

But should HITH patients be included in the hospitalisation rate? I think the data say no

TSP was right to say 5.5% is absurdly low

But it's not low because it excludes HITH. It's low because it's just plain wrong

The hospitalisation rate for in-patients alone is much higher than that

We now have visibility of this and ways to monitor it

ps. No joke, this thread took me hours to write. It's why I've been off Twitter all weekend! Monster wrap will be back tomorrow👻