A small Italian town has lost at least 1.3% of its population to Covid-19
…while the experts are speculating whether the Covid-19 fatality rate is 0.1% or 1%
The first Italian village hit by Covid-19, Castiglione d’Adda had 60 excess deaths during February and April 15th representing 1.3% of its population indicating 2% infection fatality rate (66% infection rate). Here is a story how it all happened.
One of the most crucial questions for countries deciding their Covid-19 containment strategies and actions is the actual fatality rate of the virus. Unfortunately we don’t t know it. The leading experts in the world are divided: some are arguing that the fatality rate could be 0.1%[1] or lower while some experts are arguing that the fatality rate could be higher than 1%[2]. There is a huge difference between the possible outcomes and actions that needs to be taken, depending on which fatality scenario proves to be more accurate.
In case the fatality rate is low, it means that the virus is spreading extremely fast and many people are already infected. The hospitals are burdened because this virus spreads hundreds of times faster than a seasonal influenza. All people with severe symptoms come to hospital within a month instead of coming in within the span of several months during normal flu season. Low fatality with high speed transmission would mean that heavily hit cities like New York could have most of their population infected by the end of May and life would come back to normal already in June.
In case the fatality rate is high, it means that the many countries will have long lasting, perhaps 14–18 months, fight to contain the spread and keeping the hospital burden within their capacities. Table below shows the drastic difference between alternative fatality (IFR) rates and policy actions in the case of New York.
The socio-economic consequences of these different scenarios are huge. The first one would mean quick return to normalcy and fast economic recovery would be possible, if supported by fiscal stimulus of various countries.
In the second scenario, the economy could be opened relatively soon by protecting the elderly, with rest of the people returning to work.
Finally, in the third scenario, the economy will face a huge long term pressure from long and frequent containment efforts, which will fundamentally change our society for many years to come.
Calculation of the fatality rate is simple. We just need to divide the number of people who have died due to Covid-19 with the total number of people infected[3]. To get a more precise estimate for the fatality rate, we should use the situation from 14 days[4][5] for the number of people infected and current death rate, as there is about a two weeks delay from first symptoms to death. While we can fairly accurately calculate the number of deaths by looking at the official death statistics and comparing the change from the previous year’s death statistics, we just have no idea about the total number of people infected as most of the people are either asymptomatic or only mildly sick and thus not tested.
Many countries and cities are currently implementing antibody testing to estimate what share of the population has been already infected by the virus. Those who have had the virus and have recovered should have developed antibodies in their blood. However, there is a high possibility for error in these tests arising due to various factors. The sample may not be representative and the tests can also give either false negative or false positive results. Researchers at the University of Oxford tasked with evaluating these serological tests say they’re still weeks away from solid validation and that no versions to date have performed well: “We see many false negatives (tests where no antibody is detected despite the fact we know it is there) and we also see false positives. None of the tests we have validated would meet the criteria for a good test”.[6]
Policy makers, businesses and societies should make huge long impact decisions based on the estimated fatality rate of Covid-19. Those decisions should not be based on high uncertainty tests, as the consequences of the wrongly informed decisions would be catastrophic. We need to look into real empirical numbers from regions which got affected early and observe what actually happened and how many people really died because of the virus.
Logically, we should look into Wuhan in China, the perceived origin of the virus, but we won’t be able to do so as China never reported the real statistics in regards to the number of infected individuals and fatalities. So we need to look into the next best place where the virus landed: the small village of Castiglione d’Adda, a town of about 4600 people in northern Italy.
Covid-19 case study of Castiglione d’Adda
“I diagnosed the first COVID-19 patient in Italy, which triggered the national emergency. He was a 38-year-old healthy athlete with no comorbidities. He arrived at the emergency room in Codogno on Saturday 18 February with fever, cough, and shortness of breath. He was given antibiotics but refused to stay at the hospital. He returned that evening because he couldn’t breathe and had a very high fever. He was admitted to internal medicine and given antibiotics and an oxygen mask.
On 20 February, they called me because he couldn’t breathe. He was having a CT scan: his lungs looked terrible and he had atypical pneumonia. I decided to take him to my intensive care ward and intubate him. There I talked with his wife — it was about 11:30 am — who said that two weeks ago he had attended a dinner with a colleague who, two weeks prior to the dinner, had returned from China. I suspected a coronavirus infection and immediately called my chief to request a nasopharyngeal swab. The patient did not meet the national criteria for coronavirus testing but because of his severe situation, we decided to do it anyway. At noon we sent the test to Milan.
Even before we received the test result, my decision was to quarantine the patient because the risk of not doing so was too high. I stayed in quarantine with the patient, together with the nurses, and we used PPE straight away; this quick reaction is why we were not infected.
The positive result came back at 9:00 pm. I called the chief of the hospital who declared it a crisis situation. The chief in Lombardy was contacted as were the politicians, and a national emergency was announced. Codogno hospital was put in lockdown and emergencies were sent to Lodi Hospital, which is 30 km away.”
Dr. Annalisa Malara[7]
The presumed patient zero had come from a business trip from China around 25th of January. He was asymptomatic as most of the Covid-19 cases. On 4.2 he had dinner with patient 1, Mattia. Mattia started having symptoms and within two weeks, by 18th of February he had difficulty breathing and went to the hospital. Since he had no prior travel history in China, the hospital did not strongly suspect Covid-19 as one of the possible diagnoses and was instead sent home. On the next day, he comes back as his breathing has become more difficult. Mattia is admitted to hospital.
On evening of 20th of February, Mattia’s pregnant wife and friend tested positive for Covid-19. On 21.2, Mattia’ mother and father had come to hospital and there were a total of 15 positive cases recorded while 11 towns in the province of Lodi went into lock down.
One month has passed since patient zero arrived in Northern Italy and 17 days since Mattia and patient zero had dinner. Covid-19 has had recorded daily growth of 25–35% in many places. Using 30% daily growth rate, patient zero could have already infected 10 people by the time Mattia and he had dinner. Fourteen days after Mattia came to hospital, there would have been 500 people already infected, but only 50–100 of them would become severely ill based on recent findings of asymptomatic and mildly sick people[8][9].
The lockdown that was in place was not a total lockdown of the region while the infections and transmission of the virus continued to occur. Families continued to have dinners and lunches with family members who were unaware that they were carriers of the virus and got infected. People went shopping groceries and got infected there as well. Hospitals became major centres for infections while doctors and nurses fell ill and started infecting the majority of the population. Three of the town’s four doctors became sick or were placed in self quarantine. The hospital needed to close down as one doctor can’t tend to over 100 patients. By end February, over 900 people, 20% of Castiglione d’Adda’s 4646 population, could have been infected, continuing infecting the older people in their families during the lockdown.
On the 6th of March, Costantino Pesatori, mayor of Castiglione d’Adda, said that during the last two weeks, 16 of the town’s citizens have died of Covid-19. The town had an infection fatality rate of 0.34% two weeks after the first cases were detected. On the 17th of March, Mattia’ father died. Four days later, Mattia is released from hospital after being in incubation for several weeks. Lombardia region reports 2549 deaths.
In the last week of March, the hospital of Castiglione d’Adda records that out of 60 blood donations, 40 are positive for Covid-19 antibodies indicating that 66% of the town’s population has been infected.
On 4th of April, Mattia’ wife Valentina gave birth to their daughter Giulia, a new citizen of Castiglione d’Adda.
Calculating the infection fatality rate
According to official death statistics of Italy, there were a total 71 deaths in Castiglione d’Adda[10] between February 1 and April 15, which is 60 more (+353%) than in the same period in 2019. Most of the excess deaths, if not all, were Covid-19 related. Infection fatality rate in Castiglione d’Adda would be 1.9% while more deaths may still come (49/(4646 x 66%) = 2.0%).
The people living in Castiglione d’Adda’s age structure is average by European standards. Only 17.4% of the people are over 70 years old. In Finland, for example, 15.8% are over 70 years old. Finland currently has 240 recorded Covid-19 related deaths.
While we currently don’t have data regarding the number of deaths in different age groups in Castiglione d’Adda, we can use Italy’s overall data. Only 1.1% of the deaths have been under 50 years old, 3.8% of the deaths have been of people between 50–59 years old, and most of the deaths , which is 95%, were people older than 60 years.
Using these ratios for Castiglione d’Adda: out of the 60 people who died, only 3 would have been under 60 years old, 7 would have been 60–69 years old and 50 above 70 years old. The age specific infection fatality rates would have been 0.03% for people under 49 years, 0.4% for people between 50–59 years, 1.7% for 60–69 years old, 5.9% for 70–79 years old and 14.5% for >80 years old.
What can we learn from Castiglione d’Adda?
Most probably then the virus went through almost the entire population of Castiglione d’Adda in just a few weeks during February. Although the lockdown order came on 21.2, it is probable that big part of town got infected during the February. In most of the families, there was already at least one person infected, which continued transmitting through the rest of the family even during the lockdown.
Most of the people who died were old, who got infected by the younger generation during family dinners and lunches.
Castiglione d’Adda lost in one month 2.4 times more people than on average have died in one year during 2015–2019. It is very difficult to justify a decision not to try to use all means to prevent the spread of the virus to the older community. Castiglione d’Adda tried, but it was too late. Based on the data from the town, it seems that a lockdown and strong containment strategy is only a reasonable and ethical strategy unless and until we have strong evidence of lower fatality rate.
With covid-19, the decision-makers have very little time to react. Responding a week after the first case in the community might be too late already. For example in early March Singapore implemented a quarantine to people coming into country only five day later than Hongkong. Today Hongkong has almost zero daily cases, while Singapore has over 500 daily cases as of 4.5.2020.
For people under 50 years old, the Covid-19 fatality rate is close to the seasonal flu fatality rate. Without a chance to infect the older generations, they could live and work normally. By preventing the spread to older generations, we can minimize the fatality of the virus to seasonal flu level.
How societies could response to Covid-19 in longer term
Based on the case of Castiglione d’Adda, the fatality rate of Covid-19 is so high that letting it spread freely is not an ethical solution for any society. However, long lockdowns destroy the economy and seems not feasible either. Virus spreads extremely fast and a small infected population of ten people could grow into 1000 infected cases in just two weeks. Based on this, a partial containment strategy seems inefficient and economically most expensive as full lockdowns would have to be implement again always when infection starts spreading.
Full containment strategy, where the number of infections is first contained to zero and then a continuous monitoring and alarm system is put into place, enabling the society react to potential new spread within 48 hours and start containment efforts, which focus on preventing the infection spread to older generations. Below is an example how the model could work:
1) The Government’s role should be to continuously measure, test, and alarm people. Once there is potential spread noticed, government announced red alert and society goes into pre-defined red alert mode:
2) Remote work mode for those who are older than 60 years old or who are in frequent contact with people over 60 years old or whose spouse is in frequent contact with them.
3) Government would pay remote work compensation pay to incentivise people at risk to stay at home
4) Home delivery service as the main way for +60 years to buy groceries. Potentially subsidised by the society to incentivise people to stay at home.
5) People who work in the elderly care should isolate themselves from society for four week shifts: first week: isolation and testing, second and third week working with elderly, and a week off. The amount of people working in elderly care could be double for red alert periods.
6) People under 60 years would continue working, going to school, and living normally during the virus emergency period as a way to keep society and the economy going while staying away from the older generations to keep them safe.
7) During red alert, large events and gatherings that have more than 20 people attending should be banned while restaurants should limit customers in the establishment to 1/3 of their total capacity to make enough space between tables.
8) Temperature monitoring and digital sign ups should be mandatory before people enter into public places and restaurants (Singapore and Hong Kong are some of the countries that are already doing this )
9) Usage of masks would be mandatory during red and orange alerts (Singapore and Hong Kong)
10) Once new virus cases go below the red alert threshold, say less than 4 daily cases, the red alert would be lifted , orange alert would come into force and social distancing would be recommended. Big events would still be banned, but restaurants and sport places would be open.
11) Once new cases consistently remain zero in a city for more than two weeks, the government can announce a green zone to the area and life would return to normal.
Afterword
I hope that this article was useful for the readers. We still don’t know enough about Covid-19 to be sure of its infection fatality rate. That rate can be lower and I certainly hope for it, but until we know it for sure, we should choose those decisions which with the current factual information cause least fatalities.
My deepest condolences to the people of Castiglione d’Adda and Northern Italy for your loss.
This article does not aim for scientific merit.
Antti Saarnio*
Doctor of Technology
* The author does not have background from medicine or virology and does not claim to be an expert in Covid-19
[1] CEBM 9.2.2020 “Taking account of historical experience, trends in the data, increased number of infections in the population at largest, and potential impact of misclassification of deaths gives a presumed estimate for the COVID-19 IFR somewhere between 0.1% and 0.36%”
[2] Russell et. al. Estimating the infection and case fatality ratio for coronavirus disease (COVID-19) using age-adjusted data from the outbreak on the Diamond Princess cruise ship, February 2020“We estimated that the all-age cIFR on the Diamond Princess was 1.3% (95% confidence interval (CI): 0.38–3.6)”
[3] https://www.cdc.gov/csels/dsepd/ss1978/lesson3/section3.html
[4] “We re-estimated mortality rates by dividing the number of deaths on a given day by the number of patients with confirmed COVID-19 infection 14 days before.”, Baud et. al. Real estimates of mortality following COVID-19 infection.
[5] China National Health Commission reported the details of the first 17 deaths up to 24 pm 22 Jan 2020. A study of these cases found that the median days from first symptom to death were 14
[6] https://www.research.ox.ac.uk/Article/2020-04-05-trouble-in-testing-land
[7] https://www.escardio.org/Education/COVID-19-and-Cardiology/diagnosing-the-first-covid-19-patient-in-italy-codogno
[8] https://www.bmj.com/content/369/bmj.m1375
[9] https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200306-sitrep-46-covid-19.pdf?sfvrsn=96b04adf_2
