ANALYSIS OF DEATHS FOR COVID19

Error estimation in official countings and inplications for the strategy.

Vision Paper redacted by Giulia Torrelli, Francesco Grillo and Antonio Negro.

Istat made available on October 22nd the latest update of the database which, among the many information, reports the number of deaths between March 1st and August 31st 2020 (period characterized by the COVID19 epidemic) and the average same period of the years 2015-20219 in 7.903 Italian municipalities. The comparison can be more significant than other surveys, as it is not affected by the difficult attribution of deaths due to other causes.

If, in fact, we assume [1] that the population of the same Municipality, from year to year and in the same month, does not change substantially in terms of present pathologies and in its demographic structure (for example, percentage of elderly people), it is plausible that the change in deaths is limited. This hypothesis is, moreover, confirmed by the low variance of deaths in previous years (2012 - 2018).

This means that a large and sudden variation in the number of deaths can provide an estimate of the impact of an exogenous shock such as COVID19 [2].

The Analysis

The municipalities observed by ISTAT are 7.903 and represent the totality of the Italian municipalities and they register an increase of 15% of deaths between 2020 and 2015-2019.

Looking at the numbers provided by ISTAT, an overall estimate of deaths caused by COVID in the period March 1st and August 31st 2020 is approximately 32% higher than the numbers provided by the Civil Protection (47.000 compared to 35.454 ).

Table - Death trends (2020 - 2019) and COVID19 death estimate

Tabella morti eng

Source: Vision elaboration on ISTAT and DPC date

From the table, two important elments seem to emerge:

  • The number of probable deaths (between March and August) is 1,32 (1,47 considering only the numbers in the table which, however, do not represent all Italian regions) times higher than the official one;
  • The underestimation seems lower in regions such as Emilia Romagna, Liguria, Veneto and Valle d'Aosta, while it is higher in others such as Calabria, Puglia and Sardinia.

Moreover, it is possible that the lockdown may have reduced other causes of mortality (for example, car accidents which, in general, are the first cause of death up to 50 years) and this would lead to the belief that the difference between probable and offical data is even superior. In this way it is possible to explain the trend of some regions that have seen a decrease in their number of deaths.

Therefore, it must be noted that we did not include in our analysis Abruzzo, Basilicata, Campania, Friuli Venezia Giulia, Lazio, Molise, Sicily, Tuscany, Umbrial, as in those regions there was a decrease in mortality in the period taken in consideration.

Further implications

The analysis therefore shows that there is certainly an underestimate in the official surveys of COVID-19 mortality and that this error can be so large that the ratio between real and detected deaths approaches to two.

This has further important implications both for a government involved in the management of the pandemic and the strategies adopted. Indeed:

  1. it is reasonable to imagine that the underestimation continued also in the days following August 31st and this means that the official death toll (54.363) on November 27th must be multiplied by the same error factor (1,36). This corresponds to a projection indicating a real number of victims in Italy of largely more than 70.000 (72.000). This is even more credible taking into consideration the data provided by ISTAT which refer to the monthly balance sheet on the Italian resident population. This saw, in August 2020, a decline that brought it below 60 million (59.991.186, which had not happened since 2013). Taking into consideration the trend that characterized the number of residents in Italy in the first months of the year, the trend of the same period last year and the number of deaths from COVID calculated in this paper, it was estimated (in a prudential manner) that corrently the Italian resident population could be about 59 million 890 thousand people (around 100.000 less than in August).
  2. more interesting, however, is the consequence on the number of the infection. The article published on NATURE [3] estimates - for the Chinese region of Hubei at the end of the epidemic - the real mortality rate for COVID-19 (correcting the denominator with an estimate of the asymptomatic individuals to whom the buffer has not been made) in an average value of 1,4%. Assuming that this rate may be worse in ITALY, due to the relative seniority of the population [4], this would mean that there are over 2 million and an half infected people in Italy today (around the 5% of the entire population).

In this scenario, it is possible that at the end of 2021, the Italian population will repeat a decrease similar to that recorded between January and August 2020, causing a loss of about half a million residents in two years. At this rate it is likely that in 2030 we will be closer to Spain, which in 1980 had 37 million inhabitants, than to France, which, that same year, with 55 million, had fewer citizens than Italy (56): the problem is that - after forty years - we have almost stopped and the two countries most similar to us have each added ten million inhabitants. This circumstance  poses huge problems for the Welfare and the development policies.

The implications of our projections are obviously very relevant and the next contributions of VISION in the series on the effects of the PANDEMIC are dedicated to this.

 

Bibliography

  • Baud, D., Qi, X., Nielsen-Saines, K., Musso, D., Pomar, L., & Favre, G. (2020). Real estimates of mortality following COVID-19 infection. The Lancet Infectious Diseases.
  • Colombo A. e Impicciatore R., Gli effetti della PANDEMIA da COVID19 sulla mortalità, 2020, ISTITUTO CATTANEO
  • ISTAT, Nota sull’andamento dei decessi del 2020 – Dati anticipatori sulla base del sistema ANPR, 31 Marzo 2020
  • Jung, Sung-mok, et al. "Real-time estimation of the risk of death from novel coronavirus (COVID-19) infection: Inference using exported cases." Journal of clinical medicine 9.2 (2020): 523
  • Wu, J.T., Leung, K., Bushman, M. et al. Estimating clinical severity of COVID-19 from the transmission dynamics in Wuhan, China. Nat Med (2020). https://doi.org/10.1038/s41591-020-0822-7  

 

 

[1] Jung, Sung-mok, et al. "Real-time estimation of the risk of death from novel coronavirus (COVID-19) infection: Inference using exported cases." Journal of clinical medicine 9.2 (2020): 523 Ma anche Colombo A. e Impicciatore R., Gli effetti della PANDEMIA da COVID19 sulla mortalità, 2020, ISTITUTO CATTANEO

[2] Baud, D., Qi, X., Nielsen-Saines, K., Musso, D., Pomar, L., & Favre, G. (2020). Real estimates of mortality following COVID-19 infection. The Lancet Infectious Diseases.

[3] Wu, J.T., Leung, K., Bushman, M. et al. Estimating clinical severity of COVID-19 from the transmission dynamics in Wuhan, China. Nat Med (2020). https://doi.org/10.1038/s41591-020-0822-7

[4] We assume the 2,8% which appears prudent in any case considering that the median age in CHINA - although less than the Italian one (46.5) has now grown to 38.4 years (source WORLD BANK) and that the Italian healthcare facilities however, at the beginning of the epidemic , had a greater offer.

Follow Us

Partner

vision and value logo

© 2021 Vision & Value Srl | vicolo della Penitenza 10 – 00165 Roma | P. IVA 04937201004
Credits elmweb