Risk and responsibility

As I write the world is officially in the throes of a pandemic caused by the introduction of a novel coronavirus within the human population.  This virus leads to a respiratory disease called Covid-19.  To date there have been 139,075 cases, of which 70,729 have recovered and 5,117 have died.  That gives a mortality rate of 7%.

All of this data comes from the Worldometer website. It will be different when you look.  The data may not be fully accurate, for example there may be (many) unreported cases and it is possible that not all respiratory deaths have been correctly allocated.  Error is normal in any kind of data collection, especially on this kind of scale and across multiple agents.

But this data is still extremely useful.

The data helps us to track the progress of the disease at a global scale.  The website gives country level data also in order to pull out differential rates of progress geographically, as well as differences in the number of severe cases and deaths.  This kind of information is derived from the internal properties of this data set.  If gives an indication of the scale of the problem, but may also later give an indication of the impact of any interventions.

The data also indicate a relatively high death rate.  Other sources, such as the World Health Organization (WHO), report that the risk of death is unevenly distributed across age and health categories, putting older adults and those with certain underlying conditions at greater risk.

Some people have said to me, when considering Covid-19, that tens of thousands of people die each year from influenza, thus making a relative death rate observation.  The general implication drawn appears to be that we have become used to living with that kind of risk factor on a daily basis. A further inference might be that we should not worry unduly.

It is true that large numbers of people die from influenza across the globe.  It is also the case that the data range is broad – from 250,000 to 500,000 per annum.  However, there is some evidence that the range is more likely between 290,000 and 650,000 per annum.  Moreover, those deaths are unevenly distributed both in terms of geography and age, leading to very different local death rates.

What concerns health authorities is that Covid-19 may have a much higher death rate.  WHO currently estimates it will be 3.4%, which is higher than most influenza death rates.  Whilst we have the above data we are in a very dynamic situation with various kinds of interventions to slow the spread of the disease, and the virus is itself mutating and changing.  To that end, we cannot precisely know.

But what we do know is that tens of thousands of people will die from influenza this year, and already several thousand people have died from Covid-19.  Moreover, we want to stop people dying from influenza.  There is no way of making this sound like a good situation.

The lack of clarity is compounded by our uncertainty about precise transmission mechanisms, incubation periods and chances of reinfection.  But we know enough to realize that basic hygiene measures could help enormously, as will avoidance of people in close contact with poor ventilation.

But what does help mean?  It does not mean to render us immune, but it means to delay the spread of the virus and to thus reduce the burden upon health care systems.  This will potentially reduce mortality from Covid-19, but also other diseases suffered by other patients who will be competing for access to the same health care resources.

The social distancing measures that are being discussed, and to some extent implemented, as I write are a key part of that delaying package, designed to maximize societal resource utility and reduce deaths.  What is interesting about this is that the utility is not directed at the majority of those infected.  The data indicates most people will recover and have what are described as mild symptoms.  They can manage those symptoms at home.  Thus the social inconvenience of cancelling engagements, suspending regular activities such as choir practices, coffee mornings etc., are all designed to stop infected people with mild or even no symptoms moving about, contacting one another and increasing the odds that via these transmission routes someone will give the virus to a vulnerable individual.

I will repeat this.  The more we interact the more we spread, and the more we spread the greater the chances that the vulnerable will become infected.  If that happens in a dense cluster of infections more of the vulnerable might die because resources to care for them will be inadequate.  But also more people with other conditions might die because their resources will be depleted too.  That most likely includes people with influenza.

This is an extraordinary social contract.  The request is to reduce our freedoms, our way of life, to possibly save a vulnerable and potentially large minority.  And it is a bet that we are being asked to take, because we cannot be certain how the virus will change, and how our population will respond to those changes.   But the bet, the risk, is asymmetric.

A classic example of an asymmetric risk is that of the smoke detector.  You pay a price to install a smoke detector.  But what if it is unreliable and goes off on occasions when there is no fire (say when toast is slightly overcooked)?  Should you simply do away with the smoke detector?  I hope the answer is clear – no.  Even an unreliable smoke detector is better than none because it will have a chance of detecting real fires as well as false ones.  Having no smoke detector will guarantee total failure.  We should use the unreliable one until technologies improve.  Until we can get better data.

In the case of Covid-19 and social distancing most of us are being asked to pay a price to offset the costs of others.  Those costs are potentially total – death.  The offset price is high but none the less cheaper.  But of course we all know people in those vulnerable groups, either directly as our friends and relatives, or indirectly as the friends and relatives of others.  So there is another cost here – the potential loss of valued folk.  If it turns out that social distancing cannot delay the peak of this crisis, and resources are overwhelmed, then we will have lost important social moments, which is a real problem for our kind of primate. Many of those moments could have been with our vulnerable friends and relatives.  But if we don’t do it, and it could have helped, then more people will be dead.  That strikes me as a greater sorrow.

It strikes public health officials as a greater problem, because the health service being stretched beyond capacity will require recovery in the future.  Any period of recovery will most likely continue risks for other vulnerable populations into the future, beyond the peak of this crisis.  So as a utilitarian endeavour public health policy must also think about future generations impacted by this event now.

The now famous diagram below (from Forbes under CC BY) makes this point:

So, even if social distancing seems potentially problematic that combined with hygiene measures is all we have at the moment.  We certainly should not abandon them until better technologies come along.  Most everyone will hope that those technologies will be incorporated in a vaccine, but the development of vaccines is slow, and delivering vaccination resource intensive.  In all likelihood investment in improving the technologies of social distancing and hygiene will also be worthwhile.  And any way, there will be other viruses…

Future discussions about rolling out high speed and reliable internet connectivity will perhaps be conducted within the context of public social health, and funding for end-user technology such as computers and tablets may also be considered.  Perhaps large online supermarkets will direct their corporate social responsibility efforts toward set up costs for currently excluded groups and think about technologies that might track the needs of the vulnerable before they need to order?