COVID-19 Stuff Here
- Thread starter dru
- Start date
The slowing down of the rate of increase these last few days is very encouraging, but some experts warn there may be a significant increase in the next fortnight due to recent arrivals by plane and boat testing positive. The old adage holds good here: "prepare for the worst and anything else is good news".
tragic
John Solomon (38)
It’s in the disaster plan - at least in QLD. The problem is it’s impossible to protect “clean” hospitals if this goes ballistic as patients who come in with another problem will spike a fever 3 days later, by which time half the staff and other patients on the ward will be infected. The asymptomatic infectious period and the lack of a rapid point of entry diagnostic test is what makes this virus such a problem.
And when a patient/patients in the clean hospital gets sick the public/dirty hospital is so overwhelmed it says “we can’t help - you’re on your own - deal with it yourself” so the clean hospital becomes dirty anyway.
Keeping all the COVID positive mildly sick patients who just need a little oxygen in one large temporary facility does make some sense as the staff will only need to kit up in PPE on arrival as everyone is positive - so no need to change between patients. Saves the massive use of PPE gear. If they are sick enough to require ICU they move to the main hospital. If they are improving they step down into home isolation.
Dctarget
John Eales (66)
Okay, so last night the team changed it up a lot with their modelling, they've switched from exponential to something more logistic & sigmoidal to properly represent mitigation + suppression. But I'll let them explain it their words. Sorry for the chunk of text, I need to work out how to share this stuff better. Might be best for those interested to join the facebook group.
Dctarget
John Eales (66)
Today we switch to a new model, away from the simple Exponential Growth (EG) model and move to a new logistic or sigmoidal one. We have called this model Mitigation+Suppression (M+S). We will also change from plotting the Y axis linearly, to exponentially (base 10).
This new model presents (and utilises) our full history of total case numbers, detects what phase we are currently in, and then makes projections into the future based on a small number of conditions and rates that we can adjust for each phase. Our aim is not to try and nail a prediction of the future, we simply want to illustrate how varying these rates and conditions can influence our future case load.
The new model also estimates “active cases” based on the sum of our new cases in the last 10 days (directly, no “distribution about a mean” of these cases in time). We also plot a selected percentage of active cases; presently we have chosen to display 5% of active cases. This is a conservative estimate of the proportion that might require ICU treatment. Estimates on this vary and the local data is still emerging, so interpret these numbers with care. We have refrained from providing a line representing our available ICU resources, as this is also changing with new beds and vents being brought online. For reference, in previous models and updates we have used a number of 2,000 beds (available for COVID-19) as an estimate.
If we can reduce and maintain our confirmed case growth below 3% for 3 out of 4 consecutive days, this will signal we have entered Phase 3: Suppression. We have selected a case growth rate of 1% assumed for this phase, based on an analysis of South Korea’s data (likely our best case scenario).
While we’ve made some encouraging progress this week, we want to caution everyone against overconfidence or misrepresenting these results. A sharp increase in community transmission, increased testing rates or a broadening of the criteria to be eligible for testing could all result in a break-out from the current Mitigation trend, and a return to Exponential Growth phase (see below).
We have provided 4 separate Scenarios for your consideration. It should be noted that due to the (thankfully) reducing number of cases in recent days, the differences between these scenarios are relatively minor. That would not be the case were we still in a strong growth phase (i.e. a week ago). This data should not be interpreted as an excuse to “take our foot off the pedal”. We currently have the opportunity to suppress this disease locally, we should make the very most of it. Every fraction of a percent in this rate helps us and potentially shortens the crisis.
This new model presents (and utilises) our full history of total case numbers, detects what phase we are currently in, and then makes projections into the future based on a small number of conditions and rates that we can adjust for each phase. Our aim is not to try and nail a prediction of the future, we simply want to illustrate how varying these rates and conditions can influence our future case load.
The new model also estimates “active cases” based on the sum of our new cases in the last 10 days (directly, no “distribution about a mean” of these cases in time). We also plot a selected percentage of active cases; presently we have chosen to display 5% of active cases. This is a conservative estimate of the proportion that might require ICU treatment. Estimates on this vary and the local data is still emerging, so interpret these numbers with care. We have refrained from providing a line representing our available ICU resources, as this is also changing with new beds and vents being brought online. For reference, in previous models and updates we have used a number of 2,000 beds (available for COVID-19) as an estimate.
If we can reduce and maintain our confirmed case growth below 3% for 3 out of 4 consecutive days, this will signal we have entered Phase 3: Suppression. We have selected a case growth rate of 1% assumed for this phase, based on an analysis of South Korea’s data (likely our best case scenario).
While we’ve made some encouraging progress this week, we want to caution everyone against overconfidence or misrepresenting these results. A sharp increase in community transmission, increased testing rates or a broadening of the criteria to be eligible for testing could all result in a break-out from the current Mitigation trend, and a return to Exponential Growth phase (see below).
We have provided 4 separate Scenarios for your consideration. It should be noted that due to the (thankfully) reducing number of cases in recent days, the differences between these scenarios are relatively minor. That would not be the case were we still in a strong growth phase (i.e. a week ago). This data should not be interpreted as an excuse to “take our foot off the pedal”. We currently have the opportunity to suppress this disease locally, we should make the very most of it. Every fraction of a percent in this rate helps us and potentially shortens the crisis.
Dctarget
John Eales (66)
Scenario 1: Strong Mitigation
-1.5% daily reduction in growth rate
This scenario is currently a reasonable fit for our progress since the start of mitigation which was signalled by our model on 25th March (Day 13). It is similar to the rate of reductions achieved by both China (in Hubei) and South Korea, following the implementation of full lockdowns. It is unlikely that we will be able to continue such a high daily rate of reduction, expect it to slow more gradually than currently modelled.
Dctarget
John Eales (66)
Scenario 3: Weak Mitigation
-0.75% daily reduction in growth rate
This scenario is typical of the rate of reduction achieved by Italy in the early stages of their lockdown. It might be considered typical of ineffective social distancing, isolation and quarantine measures.
-0.75% daily reduction in growth rate
This scenario is typical of the rate of reduction achieved by Italy in the early stages of their lockdown. It might be considered typical of ineffective social distancing, isolation and quarantine measures.
Dctarget
John Eales (66)
Scenario 5: No Mitigation
0.0% daily reduction in growth rate
i.e. continued exponential growth at current 3 day average rate.
The total case line on this graph is equivalent to the previous version of our model. The lines are straight here due to a shift to a log scale.
Remember, our actions today will only show up in this data in 1 week’s time (at best). The next 7-10 days of confirmed infections are likely already beyond our control.
0.0% daily reduction in growth rate
i.e. continued exponential growth at current 3 day average rate.
The total case line on this graph is equivalent to the previous version of our model. The lines are straight here due to a shift to a log scale.
Remember, our actions today will only show up in this data in 1 week’s time (at best). The next 7-10 days of confirmed infections are likely already beyond our control.
Up the Guts
Steve Williams (59)
Interesting to see what happens if/when local suppression occurs.
Keep widespread testing and the borders closed until a vaccine emerges but start opening everything back up?
Ignoto
John Thornett (49)
I guess an argument could be made that, the 'positive' trend could also be a bit of a lull due to the lower amounts of testing being done as well. Most, if not all, travelers and those who have been in contact are diminishing. It'd be great if this isn't the case, but, once the antibodies community blood testing rolls out, we'll have a better idea on the true transmission rate.
Derpus
George Gregan (70)
China has a lot to answer for.
Garry Owens
Alan Cameron (40)
Again , I want to publicly thank DC for sharing this data. Champion.
The_Brown_Hornet
John Eales (66)
Yeah me too, GO. I've been doing my own much more rudimentary analysis but the stuff that DC has done has been fantastic. Really appreciate it.
Garry Owens
Alan Cameron (40)
Well thank you anyway DC and thanks Dr Wordley
From 6.30am Monday to 6.30am this morning ( the 31st ) only a 6.50% growth in daily infection rate (266 New confirmed cases to 4,359 confirmed cases )
The next week is critical as you more fully assume of absorption of things like the Ruby Princess fuck up and non enforced quarantine of returning travellers etc - and of course , the effect of Stage 3 across the States ( 2 weeks on ) and better individual practices and habits
From 6.30am Monday to 6.30am this morning ( the 31st ) only a 6.50% growth in daily infection rate (266 New confirmed cases to 4,359 confirmed cases )
The next week is critical as you more fully assume of absorption of things like the Ruby Princess fuck up and non enforced quarantine of returning travellers etc - and of course , the effect of Stage 3 across the States ( 2 weeks on ) and better individual practices and habits