The coronavirus, mortality and life expectancy
A demographer calculates how the average life expectancy can be affected
In Sweden, we now experience the first pandemic that occurs in a society with modern information technology, and it is also the first pandemic that reaches us after having been able to follow the disease in detail for months before it began to spread in Sweden. The coronavirus (COVID-19) has already had significant effects on workplaces, shops, and our everyday life. We can all read many negative scenarios about this disease that has so far only slightly affected the health condition in Sweden - but, which is undeniable, may affect the general health condition much more. The government is therefore right to take the problems very seriously. Our hospitals may well end up under severe pressure, as experience shows from Italy and China. The coronavirus hits the elderly hard and requires a lot of resources and intensive care.
However, demographic calculations can provide valuable perspectives on how much of an impact Corona will have on mortality in Sweden in 2020. We do not know for sure how severely the disease will affect Sweden. But we can use different scenarios based on countries that have been affected in the past to calculate how it might change the average life expectancy. I have made a simple model that calculates approximately how much the impact Corona mortality can have on average life expectancy in Sweden.
A severe scenario is that we assume that Sweden will be hit roughly as hard as Italy and that Italy currently has about a quarter of the total number of deaths they will have in 2020 (in this assumption, 12,000 will die in Italy in 2020, and about 3000 have died so far). This corresponds to a scenario in which Italy continues to suffer severe damage, but soon gains control over its spread. To calculate a similar scenario for Sweden, I convert the number of deaths proportionally to the population of Sweden and Italy, and use a life expectancy table for Sweden. The calculation is based on 1500 deaths for Sweden. I use the so-called observed "fatality rate" (case of a diagnosed person dying) for Italy for different age groups and assume that the virus spreads fairly evenly between different age groups. Since Corona is more dangerous for older people, then most deaths will occur among 70-year-olds and 80-year-olds. Such a calculation would mean that the average life expectancy of 2020 would be shortened by approximately 75 days in Sweden. For a scenario where 200 people die, the abbreviation becomes about 20 days. In scenarios where the disease spreads completely uncontrolled, the number of deaths and the impact on life expectancy will be significantly higher.
The fact that the disease is highly concentrated at higher ages is an important reason why the effects on life expectancy are quite limited (in Italy, over 50% of deaths occur among eighty-year-olds and older). By comparison, smoking is assumed to reduce the average life expectancy in Sweden by about one year each year (in many other countries the effect is 2 years or more), and alcohol consumption reduces it by close to six months. 1500 deaths during one year should also be related to the death of 92,000 people each year in Sweden, and about 250 people each day.
It is natural to be afraid of a new unknown disease. However, fear and panic create many problems in society. Of course, everyone should behave responsibly and do what they can to reduce their likelihood of spreading and suffering from the virus, and follow the instructions of the authorities - hopefully, most people have realized this last week. By taking personal responsibility in combination with well thought out measures from the authorities, we can all contribute to making the situation in our hospitals manageable. But at the same time, priorities, although many people dislike talking about them, will need to be made - and then demographic calculations like the ones above can help authorities make informed choices that require careful balancing of different interests in society. It may also contribute to reduced public concern.
My calculations about Corona mortality are available in an excel document that I have posted on the Open Science Framework. It's easy to experiment with different mortality scenarios that affect the average life expectancy (including more serious scenarios than the ones I described above), and see how the deaths can be assumed to be distributed among different ages. In this document, I also discuss the model and its limitations in greater detail. It is also possible to update the model by adding information from other countries. The model is based on the assumption that nothing else in society will affect mortality but the coronavirus. If society changes in a more significant way, which seems likely, mortality rates can both increase (especially if the health care system is severely stressed) or decrease (for example, as most people are likely to live significantly healthier in the next few months) in addition to the calculations from my simplifying model.
Associate professor of demography, researcher at the Institute for Futures Studies and Stockholm University