Advertisements
Education Journal Club Outbreaks and Updates

Death Rates, Predictions and Herd Immunity

The idea of how and when our COVID19 pandemic will end is a big concern. A vaccine is at least a year away. There does not appear to be any “magic bullet” medicines at present. There are some promising studies underway but those take time. Time is something our healthcare systems do not have as they are being overrun.

This article builds on the research and work completed by W. Ridgway and a link to his presentation

In epidemiology we like to count things. We combine medical knowledge with public health data. We try and make predictions and good decisions. A lack of high-volume testing dramatically limits our data and thus, our ability to know what’s happening at present and what will happen in the future.

Looking at this morbid but important tracking method of COVID19 disease spread in communities allows us to explore several things. We are now able to have a simple formula to follow disease spread in the community based on prior knowledge of death rates. These death rates are carefully monitored and counted. Looking at deaths now gives us a window into how many in the community are likely disease carriers. Disease carriers will likely have immunity to the virus. That is currently under study but if SARS-CoV-2 behaves like other coronaviruses, there should be a period of immunity. This brings us to herd immunity.

Herd immunity is when enough people in the community have protection against the infection that it no longer spreads wildly. It’s a are thing to see an outbreak or cluster. Herd immunity begins when between 60-70% of the population has immunity.

Looking at this data, we can use death rate to give us an idea of how many in the community have had the SARS-CoV-2 virus. We can then use the death rate to monitor as we slowly expose the population, via relaxed social restrictions over a few days, to build herd immunity. Building herd immunity without overwhelming the healthcare system is critical. Sick people can go to the hospital for care. We can begin to care for those with disease other than COVID again.

I believe that a pulsatile fashion of loosening social restrictions for a few days, for certain individuals, then locking back down for a week or so will slowly build herd immunity. Medically fragile and elderly people should continue to stay at home.  Avoiding that overwhelming burn through the community is the goal. That massive surge in patients is what overwhelms hospitals. Monitoring death rates is a tracking tool to follow and decide on when to re-tighten after the relaxation.

 

Get the report in PDF: COVID19 Regional Planning PDF

COVID19 Regional Planning p1
COVID19 Regional Planning p2
COVID19 Regional Planning p3
COVID19 Regional Planning p4
COVID19 Regional Planning p5
COVID19 Regional Planning p6
COVID19 Regional Planning p7
COVID19 Regional Planning p8
COVID19 Regional Planning p09
COVID19 Regional Planning p10
COVID19 Regional Planning p11
COVID19 Regional Planning p12
COVID19 Regional Planning p13
COVID19 Regional Planning p14
COVID19 Regional Planning p15

Advertisements

3 comments

  1. Hi

    I read with interest your article “Death Rates, Predictions and Herd Immunity“ and agreed with much of it.

    However one major error which affects all your figures but which therefore has the effect of emphasising some of what your conclusions say – is your figure in the death tolls trailing indicator: the figure you use of ~14 days is incorrect.

    From Wuhan data it is reported as more like ~24 days from infection to death in fatal cases. See below from the Wuhan data. Deaths occur on average 19 days after onset of symptoms with an average of 5 days incubation to onset of symptoms.

    If you combine this with two other facts – the conclusions in your article become more pronounced

    1) ~20% to ~30% of infected are asymptomatic meaning mortality rates may be lower than generally predicted and therefore the interpolated infected 24 days before death much higher.

    So for example 5,000 deaths in a community gives an interpolated infected at onset 24 days prior (using the WHO CFR of 3.4%) so gives a figure of 147,058 infected which with a 3 day doubling time means x 256 = 37,646,000 infected in the community at time of death. (Assuming no lockdown.)

    However at a 1% CFR (which is the given Wuhan CFR) gives 500,000 x 256 = 128,000,000 infected at time of death.

    However if interpolating from the figure suggested of a 34% asymptomatic untested infected rate so we reduce CFR to say 0.66% then we get from that 5,000 to 757,576 x 256 = ~194,000,000 Community Infected at time of death.
    (See for example CDC recently on the asymptomatic.)

    2) Official death rates are based on only officially positively PCR tested deaths and therefore exclude many community deaths or even hospital deaths not recorded in the official WHO Covid-19 statistics. See for example Wuhan, UK, US and German data. This too means the mortality figures are underestimated.

    Warm wishes

    1. Bill,

      Thanks for that reply, I’m honored you took the time.

      You asked several complex questions – I will tackle one of these here:

      Where did the significance of 14 days come from?

      I came at this number two ways. First, several studies started leading me towards this number:
      https://www.worldometers.info/coronavirus/coronavirus-death-rate/#days

      But then it was actually seen in the wild, in large volume, based on numbers from Italy. I have updated the pdf to make this more apparent. In the yellow area you will see the projection of 14 days resulting in a big downward trend as that window extends further.

      You can access the data as well as projections for NY and Louisiana here:
      https://docs.google.com/spreadsheets/d/1aTT1gG1AB-gMi6mnRC79nJQ1kGYOD2In94qDtzsyHZg/edit?usp=sharing

      If you click on the Italy tab and look at cell E36 you will see the single biggest jump is on day 14-15 after Italy began its lockdown. This is a sample set of hundred of people dying per day, the largest sample set I’ve seen, pointing to the significance of the 14 day interval.

      Note I live update that version, so it will change.

      Wyeth

Leave a Reply

%d bloggers like this: