One Finite Planet

One Finite Planet

New York, Sweden and the case for herd immunity.

Date Published:

This post examines the idea that achieving herd immunity from an outbreak of Coiv-19 just might be a solid strategy. Not ‘let it rip’ herd immunity, which I will also review, but managed and optimised herd immunity where death and serious illness is kept to the lowest level possible. Two locations with that may have sufficient herd immunity already to impact their strategy going forward are New York and Sweden, and both of these are considered.

  • Let It Rip
  • Optimisation – And the Assumptions
  • New York
  • Sweden
  • Conclusion

Let It Rip

Sometimes the term ‘herd immunity’ is taken to mean: “rely on herd immunity alone”. In 1918, some areas the outbreak came to end solely as a consequence of herd immunity. Why not with Covid-19?

The problem with what I am calling ‘let it rip’ which could be described as ‘herd immunity without flattening the curve’, is that the death toll would be unnecessarily high. It could be expected to have at least double the number of deaths, since a fully functioning medical system is able to save at least half of those who would die without treatment, and a totally overloaded medical system would result in no treatment being available for most cases.

Consider the situation in New York where it is considered, that medical systems reached their limit during the March ‘peak’. Even assuming some herd immunity, removing lockdowns would still see a medical system completely overloaded.

I have seen statements that lockdowns are really about creating hardship for the younger people in order to protect older people. However, with measures to contain an outbreak, the number of younger people who also die is sufficient to also provided motivation to younger people. It is only with a successful lockdown that the number of younger people dying becomes of low significance.

In short, the only advocates for ‘let it rip’ I have heard, are not accurately considering the resulting outcome with Covid-19.

Optimisation – The Assumptions

‘Let it Rip’ is not very desirable, but there is an alternative path to achieving herd immunity: and optimised outbreak. An optimised outbreak is one where sufficient people to achieve herd immunity are infected, but:

  • case loads are kept within the capacity of the medical system
  • the death toll can be kept much lower by selection of those who are infected

The strategy to deliver the first of these goals is well known: slow the spread by ‘social distancing’ and lockdown measures.

It is the ability to achieve the second goal that is unproven at this time.

Overall, ‘optimised herd immunity’ relys on the three assumptions:

  1. younger people are significantly less vulnerable and have far lower mortality rates and cases of severe illness
  2. it is possible to weight cases such as to protect vulnerable people
  3. children are almost entirely unaffected
  4. infected people remain immune for a sufficient period for herd immunity to result in effective eradication
  5. immunity as a result of an outbreak occurs prior to a vaccine or cure

1. Younger People are Less Vulnerable

Some of the most accurate estimates available suggest that the true mortality rate for Covid-19 is approximately 0.5%, as most figures, including the data for this table provided here, are relative to confirmed cases, which is for lower number than actual cases.

As quoted on Business Insider

It is clear that if those aged 75+ could have been fully protected from infection, the number of deaths would have been approximately halved.

What I have not found for purposes of comparison, is a chart for hospitalisations and deaths prior to Covid-19. Note that, for example, the death rare from Covid19 was 2.2 times higher than for the 65-74 age group (25.57% vs 11.51%). However, the death rate from all medical reasons would be expected to be higher for the older age group. In any normal 3 month period, what percentage of people 75+ die relative to those in other groups? Clearly there are more deaths and hospitalisations than normal, but perhaps the risk proportionate to the expected risk of medical problems for each age group?

Note also, that the number of confirmed cases for the 0-17 year age group is almost 20 times lower than the confirmed cases for 18-44 age group. The same number of cases could result in 60 deaths from this age group. What is not certain is why this number is so much lower. Are there far more asymptomatic cases? Are younger people far less likely to be infected? Was this age group better isolated from the sources of infections?

It may also be interesting to consider ‘lost years’. The average life expectancy in New York at birth is 81.2 years. Each year you have already survived slightly increases your life expectancy, so not all ages actually have the same life expectancy. I do not have an age age adjusted table, but as premature death (before age 65) is at less than 0.2% from the same government data source, only a small adjustment would occur for those younger than 65. Beyond the age 65, more data is required for accuracy.

However, typically:

  • each 0-17 year old never was able to have any adult life at all and lost over 90% of their life expectancy
  • each 18-44 year old most likely was able to get married and have children, and would be leaving young children without a parent and lost 60% of their life expectancy
  • each 45-64 year old would likely have older children, and would on average around 1/3 of their life expectancy.

In summary, younger do appear less vulnerable, what is in question is how much less vulnerable. Then comes the difficult moral question as to as to whether a life lost at a younger age is overall a greater ‘cost’ than an older person who had the chance to experience more of their life losing their remaining years.

From this data, it is not clear exactly how much less vulnerable younger people are, and it is clear they are not invulnerable.

2. It is possible to protect vulnerable people

Younger people certainly far less vulnerable than the elderly. If it is possible to full quarantine the elderly, then perhaps mortality rates could be halved? Further, even deaths amongst those not in the elderly group those with underlying medical conditions represent are overrepresented in statistics. Perhaps further quarantine for these groups?

Consider that Sweden, a country with a strategy that can arguably be considered as targeting an optimised herd immunity strategy, still have the majority of deaths occurring in the elderly age group. If fact it is reported that Sweden has over 1700 if approximately 2000 deaths (as of April 24th) having been in the over 70 age group. Despite specific measures to protect that group, the statistics for Sweden at least appear worse than those for New York, where lockdowns are universal, not just lockdowns targeting protecting the elderly.

3. Children are almost entirely unaffected

I consider this one of the most dangerous assumptions. There have been suggestions that children, who typically account for around 25% of the population could be infected almost zero real cost. This assumption can play a role in a variety of immunity strategies, and can be one motivation for wanting children mixing in schools and at day care centres even if parents are working from home. I will explore this further in a future post.

4. Infected People Remain Immune for Sufficient Time

A big question with Covid-19 is how long people remain immune. This article quotes leading experts.

Consider Influenza. People need a new vaccination every year. For the common cold, immunity is even shorter with multiple infections in a given year common. However, for Polio and Measles, immunity is often life long. What about Covid-19? The short answer is, it is too early to be certain.

Any strategy of herd immunity has to factor in the risk that herd immunity may only last:

  • Months: this would render a herd immunity strategy useless
    • Even if such short immunity is of low probability, it is possible.
  • One Year: an outbreak could be expected to result in eradication
    • The same outcome as ‘eradication’ (but with more deaths) and with borders needing to close prior to immunity expiring
  • Three Years or More: Worth the cost?
    • If there is no vaccine/cure then ‘short term pain for long term gain’
    • The result would be not just eradication, but ongoing protection even with open borders

5. There is no early vaccine or cure/treatment

A vaccine also delivers herd immunity, without the medical cases, severe illness and the deaths. A cure or treatment could mean cases could avoid much of the medical system loading, and most of the illnesses or deaths.

If either of these solutions is available in reasonable time frame, then the cost / benefit equation of a managed outbreak would be become difficult

New York

New York now has some data from random testing from public locations. This data suggests 13.9

Preliminary data shows about 13.9 percent of the population of New York state — about 2.7 million people — have at some point been infected with the coronavirus.

About 3,000 people were randomly tested at grocery stores and other public locations to allow officials to get a broader sense of how widely the virus has spread in New York and how many people might now have immunity.

In New York City, 21 percent of residents had antibodies for coronavirus, compared with 3.6 percent in upstate New York, 16.7 percent in Long Island and 11.7 percent in the Westchester and Rockland area.

“They were infected three weeks ago, four weeks ago, five weeks ago, six weeks ago. But they had the virus, they developed the antibodies and they are now recovered,” Gov. Andrew Cuomo (D) said Thursday.

From ‘the hill’ April 23.

It is important to note that the data is not completely random, as those tested were at grocery stores and public locations, which means anyone who never leaves home would not be tested. However, if the figures are representative, New York city itself would be have 20% immunity. This would be enough to impact the virus spread, slightly lowering the contribution to lowering the spread required from social distancing. It also means that the city itself is almost halfway towards the 50% level generally quoted as required to reach herd immunity.

While there are two important points:

  • not including people who do not leave their homes may result in inflated including positive results
  • higher numbers for New York city may also represent a higher spread rate in the City than elsewhere in the state

This if correct means 14.9% of 19.45 million = 2,820,250 people infected resulting approximately to 14.9% of being infected, with state deaths at approximately 16,000 deaths. A mortality rate of approximately 0.6%. This means a 50% infection level for the entire state would be approximately 55,000 deaths. An additional 39,000 deaths, unless some system of isolating the most vulnerable could be introduced from now on.


Does Sweden have a lockdown?

Is Sweden targeting a ‘herd immunity’ outbreak?

The answer to both these questions could be ‘yes’ or ‘no’, depending on how you interpret the question, and both answers can be found depending on what you read.

Firstly, on the lockdown. Sweden has not introduced a lockdown by introducing laws, but instead by providing information on what behaviour is expected. Given Swedes are generally civil minded, this approach can be very effective. A slightly lower number of people switched to working from home than in neighbouring countries, but then Sweden already has more people working from home than in those neighbouring countries. Businesses did shutdown, and are changing business processes before they are to reopen. However, social distancing guidelines set by the government are more relaxed than for other countries. Guidelines limit gatherings to 50, not to just two. If voluntary guidelines were as strict as laws in other countries, then using voluntary guidelines in Sweden would likely have the almost the same effect as legally enforced lockdowns in other countries. So Sweden has a voluntary lockdown, but also a more relaxed lockdown than neighbouring countries.

Secondly, on Sweden having a herd immunity outbreak. Sweden does have lockdown measures, so there is no ‘let it rip’ approach that some people associate with herd immunity. But is the target to reach herd immunity through an outbreak. That does not appear correct either. There are reports Stockholm may be close to herd immunity level, but the same reports state that this will occur at the time of peak new cases. If this is true, then this is peak cases under lockdown. Herd immunity under lockdown. New York, Italy and everywhere else has also reached herd immunity while under lockdown‘. So no, Sweden is not really on a herd immunity strategy.

The argument Sweden can claim is that ‘while we only have herd immunity while lockdown like everyone else, this is a lockdown we can live with‘.

… authorities have closed senior high schools and banned gatherings of more than 50 people. …Statistics show roughly half the Swedish workforce is now working from home, public transport usage has fallen by 50% in Stockholm and the capital’s streets are about 70% less busy than usual – but Swedes are still able to shop, go to restaurants, get haircuts and send children under 16 to class even if a family member is ill.

The Guardian, April 15.

Perhaps it is liveable, perhaps even a new normal, but is the streets are 70% less busy, then it is not ‘life outside of a lockdown’ either.

So, how many people are now immune? This is a subject of debate, while there is a suggestion that almost 50% of people in Stockholm are now immune, there is also data that casts that into doubt.

Mathematical models have also been performed to estimate the community spread of SARS-CoV-2. In analyses conducted by a leading UK group, 3.1% of the Swedish population was estimated to be infected by March 28. This contrasts with the much higher proportions estimated for Stockholm by Tom Britton, a leading Swedish academic working with Folkhälsomyndigheten, who suggests up to half of the capital’s population will be infected by the beginning of May – and the rest of the country may follow suit quickly.

But how can you get such different estimates?

Real Clear Science April 27.

Not only do the statistics quoted above question genuine herd immunity being close, here is another simple calculation. New York has twice the population of Sweden. Calculations for New York, where they have conducted random sampling (unlike Sweden at this time), suggest less than 14% of people have been infected, and that is after 16,000 deaths. If health systems perform at the same level, Sweden would have half (because there are 10 million Swedes and 20 million in New York) the deaths at 8,000 to produce 14% of people immune. In fact Sweden has around 1/4 of that 8,000 equivalent number of deaths, and yet there is a suggestion those fewer deaths are from over 3 times more cases in Sweden? Whilst this may be possible, reasons for such a difference of 12 twelve times less deaths per Covid-19 infection are not explained anywhere I have seen.

So Sweden may be headed towards outbreak based herd immunity, but it would seem more likely based on available data that herd immunity may also be first delivered through a vaccine.


There are two key lessons from the 1918 flu epidemic and the famous comparison between Philadelphia and St Louis. Firstly, as with Philadelphia, herd immunity through an unmanaged outbreak comes at an enormous cost. Secondly, as with St Louis, a peak under lockdown is not yet herd immunity. Countries have been very successful are creating ‘lockdown peaks’ with Covid-19, but no country is yet close to a herd immunity peak.

Projections by New York modelling as at April 24

Data so far suggests New York may be closest to herd immunity at their own lockdown peak, but on their own projections, the path to reaching herd immunity is still a long one.

While questions remain on how long immunity lasts following infection, it would seem too early to yet target herd immunity through an outbreak. The best hope for herd immunity at this time is a vaccine.

Table of Contents


Covid-19 & Vaccination Deaths: Statistically, Coincidences will distort reported deaths.

I read recently about reasonable people protesting over post vaccination deaths in South Korea, echoing stories from around the globe about the underreporting of deaths following vaccination.

Can most of these deaths be just coincidences? This question has me seeking the real story on what is happening, not just with deaths following vaccination, but also with deaths from the virus. Almost one year after my initial exploration of vaccine efficacy and safety, now there is data, not just projections, so it is time for a review, and this question needs answering for any such a review.

Read More »

Covid-19 Immunity: Layers, immunity from vaccination vs infection, and the good news.

Shrek: No. Layers. Onions have layers. Ogres have layers. Onions have layers. You get it? We both have layers.


Instead of a single immune response, the human immune system can be though of as having several layers of immune response, that all fade at different rates.
Immunity is a complex set of different layers all subject to change over time at varying rates, making comparison between immunity following vaccination or infection highly time dependant.

Read More »