On June 26, 2020, the number of cumulative deaths attributed to the Covid-19 epidemic in France was 29,778.
And the number of daily deaths continued to decrease with 26 new deaths recorded in 24 hours at the hospital. This steady decline in the number of daily deaths indicates that France is approaching its final Covid-19 death toll in 2020. In the absence of a second wave, it should stabilize around 30,000 death.
What will be the consequences of this unprecedented epidemic on life expectancy in 2020? Beyond the Covid-19 death tally, which is essential for monitoring the progression of the epidemic, the impact in terms of life expectancy makes it possible to place these deaths in the more general context of French demography. The total number of deaths attributed to the Covid-19 being specified, it is possible to make initial estimates.
Relatively modest decline in life expectancy
Comparing the daily deaths of 2020 to those of 2019 over the same period, we observe that the Covid-19 epidemic generated a strong excess mortality concentrated in the months of March and April.
In early April, at the peak of the epidemic, there were in fact up to 1,000 daily deaths in excess of 2019, an increase of more than 60%. Since the beginning of May, however, the number of daily deaths has returned to a level comparable to that of 2019. This indicates that even if the epidemic continues to claim victims, it has weakened to the point of no longer generating a surplus visible of death compared to 2019.
These observations allow us to begin to understand the impact of the Covid-19 epidemic on life expectancy in France in 2020. It is possible, for example, to assume that daily deaths on the rest of the year 2020 will remain similar to the deaths of 2019, as has already been the case for several weeks.
This assumption is quite crude, but it nevertheless allows preliminary estimates to be made. Indeed, apart from major health events, the number of daily deaths varies little from one year to another.
Our results indicate that the decline in life expectancy in France in 2020 should be relatively modest. For men, this would be a decrease of 0.2 years, representing a life expectancy of 79.5 years in 2020 compared to 79.7 years in 2019. For women, life expectancy is expected to decrease by 0.1 year, or 85.5 years in 2020 compared to 85.6 years in 2019. As for life expectancy at 65, with 0.3 years of decline for men and 0.2 years for women, the losses should be a little more marked, but ultimately quite similar to those of life expectancy at birth. The greater loss of life expectancy among men reflects their greater vulnerability to this disease.
What do these decreases in life expectancy represent? Even though life expectancy has increased by an annual average of 0.2 years in recent decades, such declines are far from unusual. For men, a similar decline occurred in 2015. For women, at least as large decreases occurred in 2008, 2012 and 2015. These episodic decreases are mainly explained by variations in intensity and timing of seasonal flu from year to year.
Another point of comparison is the heat wave episode in 2003. In that year, life expectancy fell by 0.1 years for women, but increased by 0.1 years for men. The declines expected in 2020 are therefore not exceptional.
Why such a modest drop?
It is undeniable that the containment measures made it possible to decrease sharply the number of deaths due to Covid-19. A study estimated that these exceptional measures have reduced the number of Covid-19 deaths by 83%. Without containment, the expected decline in life expectancy would certainly have been much greater.
The expected loss of life expectancy estimated here, however, appears modest in light of the nearly 30,000 Covid-19 deaths that have occurred to date, despite the containment measures. Several explanations can be put forward. First, it should be noted that deaths at older ages have a less significant impact on life expectancy than deaths at young ages, since deaths at older ages result in fewer lost years of life. However, the vast majority of Covid-19 deaths occurred at relatively old ages (82% over 70).
Second, the number of excess deaths in January-May 2020 compared to 2019 over the same period is ultimately lower than what one would expect given the Covid-19 death toll. In fact, there were around 287,000 all-cause deaths (Covid-19 or not) between January and May 2020, compared to the 270,000 all-cause deaths recorded over the same period in 2019. The excess of deaths in 2020 is so around 17,000 deaths, or 12,000 deaths less compared to the 29,000 Covid-19 deaths recorded over the same period. This means that compared to 2019, there was a deficit of 12,000 deaths in January-May 2020 from causes other than Covid-19.
This deficit of 12,000 deaths outside Covid-19 can be explained by three reasons. First, the year 2020 started in a particularly favorable way from the point of view of mortality. There were in fact approximately 10,000 fewer deaths in January-February 2020 compared to 2019 over the same period, this due to a seasonal flu epidemic that did not cause significant mortality unlike those of four of the previous five winters, including that of 2018-19. These death deficits in January-February partially offset the excess deaths in March-April due to the Covid-19.
Second, some of the victims of Covid-19 may have been frail people who would have died between March and May 2020 even in the absence of the epidemic. The excess of deaths due to Covid-19 could therefore lead to a drop in deaths from other causes and contribute to the observed deficit of deaths outside Covid-19. These effects, called “Harvest effects”, had explained in part the rapid decline in deaths from the 2003 heat wave, and similar effects can be expected with the Covid-19 epidemic.
Finally, it should be noted the “indirect” effects of containment measures, which, by profoundly changing the lifestyle of French people, could have resulted in a decrease in deaths due to causes other than the Covid-19, for example the road accidents.
In total, the 12,000 non-Covid-19 deaths averted between January and May 2020 mitigate the impact of the epidemic on life expectancy. Without these avoided deaths, the expected decline in life expectancy in 2020 would have been approximately 0.5 years, which would have represented a much more unusual loss.
Uncertainties for the rest of 2020
Our estimates of the expected decline in life expectancy in 2020 are based on the assumption that deaths over the rest of 2020 will be similar to deaths in 2019 over the same period. They do not take into account the possibility of a second wave of the epidemic by the end of 2020. It is however unlikely that a second wave, if it takes place, will be as lethal as the first because of better knowledge of the virus and of the best preparation of health authorities and the population for this epidemic.
Our estimates also do not take into account the possibility of an increase in deaths related to waivers or postponements of care during the confinement period. For cancers, for example, delays in screening or treatment could lead to additional deaths by the end of 2020. But again, it is unlikely that this type of effect is large enough to significantly affect our results.
Conversely, our estimates do not take into account the possibility that daily deaths over the rest of 2020 will be lower than those observed in 2019 over the same period, for example due to possible harvesting effects. This could further mitigate the decline in life expectancy in 2020 and even generate a slight increase, as had been observed for men in 2003 despite the heat wave.
There is therefore obviously some uncertainty surrounding the drop in life expectancy in 2020, and it will be necessary to wait until the end of the year for a definitive answer on this question. Despite these uncertainties, data to date indicate that due to multiple factors, the Covid-19 epidemic is not expected to generate a particularly unusual decline in life expectancy in France in 2020.
Michel Guillot, Research Director, National Institute for Demographic Studies (INED) & Population Studies Center, University of Pennsylvania, National Institute for Demographic Studies (INED) and Myriam khlat, Research director, National Institute for Demographic Studies (INED)