Here's the graph for Wednesday 15th April from Chris McDowall | The Spinoff

Gudbjartsson et al (2020), Spread of SARS-CoV-2 in the Icelandic Population, New England Journal of Medicine, 14 April

https://www.nejm.org/doi/full/10.1056/NEJMoa2006100?query=RP

(only skim read)

Here's the graph for Thursday 16th April from Chris McDowall | The Spinoff

The flattening out of the bottom ("active" cases) is a tad worrying, I'd hope for a fairly steep continuing decaline.
Maybe tomorrow's figures will bend the graph back in the 'right' direction. 🤞

Looking at this graph, the blue "active" line parallels the "new cases", just 2 weeks to the right. Active cases is, if you like, the integral of the new cases, summed over the previous two weeks. (Yes, it's an approximation and a simplification, but I think it's near enough). Right now, the tail end of the current 2 week window is moving across the flat part (1st April or so) of the "new cases" graph, so the reduction in active numbers is due to finding less and less new cases. So in the next couple of days, as the tail of the window moves to the descending part of the "new cases" graph, I expect the "active" line to start dropping fast. :)

PolicyGuy:

 

Here's the graph for Thursday 16th April from Chris McDowall | The Spinoff

 

 

 

 

The flattening out of the bottom ("active" cases) is a tad worrying, I'd hope for a fairly steep continuing decaline.
Maybe tomorrow's figures will bend the graph back in the 'right' direction. 🤞

 

How much of the steepness is due to lower testing due to easter?

frankv:

 

 

Looking at this graph, the blue "active" line parallels the "new cases", just 2 weeks to the right. Active cases is, if you like, the integral of the new cases, summed over the previous two weeks. (Yes, it's an approximation and a simplification, but I think it's near enough). Right now, the tail end of the current 2 week window is moving across the flat part (1st April or so) of the "new cases" graph, so the reduction in active numbers is due to finding less and less new cases. So in the next couple of days, as the tail of the window moves to the descending part of the "new cases" graph, I expect the "active" line to start dropping fast. :)

 

Which makes it more concerning about thinking about exiting lockdown this time next week (even to L3). there are too many infected people to open things back up.. ideally the blue line is basically zero if we are going to eradicate... before we exit L4 lockdown...

Confirming my suspicions that this is a manual process at the MoH, the format of the downloaded spreadsheet today is different from previously... the date columns are now labels, not numbers. :( Nevertheless, I've converted back to numbers for my graph.

I'm not concerned about the slight uptick at the end of the 7-day average curve. I think that's just an artifact of the averaging, predicting a linear trend for the next 3 days based on the last 3 days, and a slight increase in numbers due to increased testing results after Easter.

frankv: Active cases aren't a problem if they're in quarantine, which all the recent imported cases are. I expect that mandatory quarantine for all confirmed/probable cases is in our future.

Will need to be careful that the possibility of being forced into quarantine does not have a negative impact - with people knowing they have symptoms not wanting to be tested.

Fred99:

 

 

 

Will need to be careful that the possibility of being forced into quarantine does not have a negative impact - with people knowing they have symptoms not wanting to be tested.

 

Thats a good example of why the sentinel testing is being kept quiet. Some people may be paranoid, and are super safe in their bubble due to that, and will rush to get a test, meaning a group of super low risk people skew the results. And those that "dont want to know" dont get tested. But are more likely to at a test site at the supermarket, where they will get comforted and reassured.

Edit

Barry said they did not advertise the testing centre earlier as this would bring in the wrong target group. They were looking for people with no symptoms or known exposure to the virus. 

tdgeek:

 

Fred99:

 

 

 

Will need to be careful that the possibility of being forced into quarantine does not have a negative impact - with people knowing they have symptoms not wanting to be tested.

 

 

Thats a good example of why the sentinel testing is being kept quiet. Some people may be paranoid, and are super safe in their bubble due to that, and will rush to get a test, meaning a group of super low risk people skew the results. And those that "dont want to know" dont get tested. But are more likely to at a test site at the supermarket, where they will get comforted and reassured.

 

Maybe someone here who likes stats could make a stab about significance of getting a single positive result from the 1000 or so supermarket tests being carried out.  

While I don't think it's a waste of time, the chances of finding positives are also hindered by false negatives in testing, especially in pre-symptomatic or asymptomatic cases who may not have active infection in the nasal cavity when tested. I suspect that we'd need to do tens of thousands of tests to get meaningful data to quantify undiagnosed cases.  But OTOH if we find one in the thousand or so already tested - then we've got a problem and should probably abort plans to scale down from level 4 - until more is known.

Fred99:

 

 

 

Maybe someone here who likes stats could make a stab about significance of getting a single positive result from the 1000 or so supermarket tests being carried out.  

 

While I don't think it's a waste of time, the chances of finding positives are also hindered by false negatives in testing, especially in pre-symptomatic or asymptomatic cases who may not have active infection in the nasal cavity when tested. I suspect that we'd need to do tens of thousands of tests to get meaningful data to quantify undiagnosed cases.  But OTOH if we find one in the thousand or so already tested - then we've got a problem and should probably abort plans to scale down from level 4 - until more is known.

 

There is that, you know more than me about the techomilogical stuff, but they seem to feel that its worth it. If they stick to supermarkets which I imagine they will as todays its PakNSave Moorhouse, testing those employees might be useful. Has any supermarket customer facing person been infected?

Hopefully someone may ask about a or pre symptomatic testing and how reliable the results are tomorrow. 

Here's the graph for Friday 17th April from Chris McDowall | The Spinoff

The "Active cases" trend is comfortingly trending solidly downwards

Update of plot comparing total cases per million population for NZ, Australia, South Korea, Hong Kong & Taiwan. 

BUT ... be aware that our apparently low figure yesterday (8 cases) appears to be an illusion.  It is mainly due to ~half of yesterday's cases being already included in the figures released yesterday.  That is not normally the case; eg. no cases confirmed today are in today's data release.  Will post a follow up plot showing this shortly.