NZ COVID-19. Conclusions at initial Lockdown midpoint. NZ should stay at Level 4 Lockdown for an extra 4 weeks.

Yesterday marked the midpoint for New Zealand’s initial four week Level 4 Lockdown period. This period still has another two weeks until it ends. The Government has stated that it will not end early.

Below are my conclusions and a copy of an email I sent yesterday to some Party leaders and the Director-General of Health.

We have concluded that New Zealand should stay at Level 4 Lockdown for an extra 4 weeks.

For a detailed analysis please my previous posts:
https://aaamazingphoenix.wordpress.com/2020/04/07/covid-19-nz-ii-how-many-cases-and-deaths-may-we-expect-when-should-nz-come-out-of-level-4-lockdown/
https://aaamazingphoenix.wordpress.com/2020/04/06/covid-19-nz-how-many-cases-and-deaths-may-we-expect/

The amount of money the NZ Government has spent on an online learning package ($87 million) strongly indicates that students are unlikely to be at school for some time after the end of the initial Lockdown period.

New Zealand is doing extremely well with its COVID-19 outbreak.

We still only have one death.

The cumulative number of cases has not become exponential since 24 March (Day 26).

Since 24 March the cumulative number of cases has stayed essentially linear.

All our estimates for the total number of cases exclude a very long tail (once numbers become very low) because as China has proven a very low tail can be very long (over one month). We allow 10% for the tail but China has shown this may not be enough.

In previous posts we estimated a case range from 1,350 to 2,700 (excluding any long tail).

In case there was a long tail, we added on 10% to get a range from 1,500 to 3,000 cases.

The ranges were produced both by looking at graphs for New Zealand (alone) and also by looking also at the data for Australia and Norway.

By my calculations, 5 April (Day 38) is likely to be the midpoint for our COVID-19 outbreak.  Even if the Government thinks the midpoint may have happened sooner, it would be safer to assume 5 April as the midpoint.

The lower limit for the range was also obtained by assuming that the midpoint may be around 31 March and 1 April (Day 33 and Day 34).

On 5 April (Day 38) we had 1,039 cases. This appears to be the midpoint.

On 6 April we had 1,106. The sum of these is 2,145 cases, an estimation for the total number of cases. We can add on 10% for a long tail to get 2,359 cases which we round to 2400.

On 2 April (Day 35) we observed the number of cases still appeared to be roughly linear (a straight line) so we used least squares (linear regression) in Excel to get an equation for the line:

y=70.65x – 1681.27 (2dp)

where x is the Day number.

Apart from our assumed midpoint (5 April) and the next day (Day 38 and Day 39), the numbers have followed close to this line.

Yesterday (8 April; Day 41) we reached 1,210 cases. The line estimate was 1205, very close.
On 24 March (Day 26) we had a total of 155 cases. The line estimate was the same (truncated).

The Lockdown has kept these cases linear.

Up until yesterday there has been no flattening of the curve.

From today we should see a new line estimate soon with a slope which is less steep. i.e. finally the curve should flatten starting today.

We hope that Lockdown will prevent another peak.

The Day number is used to estimate the number of cases.

The first case actually occurred on 28 February (Day 1).

The midpoint April 5 (Day 38) is the halfway mark.

Most cases will have appeared by Day 76 (2 x 38).

If we are lucky and the curve flattens all new cases may have stopped before 38 days after the midpoint i.e. Before Day 76 (13 May).

We should extend the Lockdown an extra 4 weeks to just before 28 May to allow for a long tail and to make sure there are no new infections.

Below is a copy of an email message I sent yesterday.

I will send copies to other Party leaders today.

I will also send a copy to the Minister of Health. I understand he may be incredibly busy trying to “remember” the other occasions when he has breached Lockdown.

===========================================================

——– Forwarded Message ——–

Subject: 5 April was likely to be the NZ COVID-19 midpoint. NZ should not come out of Level 4 Lockdown for 4 extra weeks!
Date: Wed, 8 Apr 2020 15:47:04 +1200
From: Alan Grace
To: ashley.bloomfield@health.govt.nz
CC: Alan Grace, jacinda.ardern@parliament.govt.nz, simon.bridges@parliament.govt.nz, winston.peters@parliament.govt.nz

Dear Dr Bloomfield

I realise that you are probably familiar with the information I am providing.

Nevertheless I hope that some of my results may prove useful to you and the Government in your decision-making.

Further to my email yesterday, the graphs I included in that email and this email indicate that 5 April is likely to be the NZ COVID-19 case midpoint.

Public statements from you some days ago indicate that the midpoint may have been thought to have occurred earlier (this may still be the case).

The red straight line in the graphs remain a good indicator today.

This may be surprising since extrapolation (estimation of values beyond the original range) can be unreliable.

For today, the total number of cases was 1210, the red line produced 1215.

The red line was calculated on 2 April in Excel using linear least squares estimation using data from Day 25 to 35 (23 March to 2 April).

Today is Day 41 (8 April). Day 1 was for my calculations was 28 February.

Please see today’s graphs below and:

COVID-19: Proof that 5 April was NZ COVID-19 midpoint. NZ should not come out of Level 4 Lockdown for 4 extra weeks!
https://aaamazingphoenix.wordpress.com/2020/04/07/covid-19-nz-ii-how-many-cases-and-deaths-may-we-expect-when-should-nz-come-out-of-level-4-lockdown/

Below are updated graphs including today’s figure:

NZCases8Apr

BarChartCases8

Regards

Alan Grace

M.A. hons (Mathematics)

===========================================================

5 Comments Add yours

  1. Billy says:

    There is very good modelling that suggests that measures other than level 4 lockdown would have produced very good results and very few “deaths prior to the appointed time”

    Click to access Corona.pdf

    Liked by 1 person

    1. AlanGraceNZ says:

      Thanks, I had not seen this report. I see that it assumes an infectivity period of 10-11 days rather than usually assumed 6 days (range 2-11). I am not sure if you have seen the MOH reports:
      https://aaamazingphoenix.wordpress.com/2020/04/01/covid-19-in-nz-did-you-think-my-predictions-were-bad-what-do-you-think-of-these-scenarios-modelling/

      Like

Leave a comment