Case Fatality Rate in Italy – Reasons for Overestimation

Readers may start seeing news headlines about alarmingly high mortality rate in Italy. I want to explain why we think this number may (hopefully) be an overestimation of the true fatality rate.

Let us first define what the case fatality rate (CFR) is:

CFR = number of deaths with COVID-19 / number of COVID-19 cases

The newest reported case fatality rate in Italy is now 7.2%. This is an alarming number, more than double what many other countries have reported (for comparison China’s CFR is 2.3%).

We believe there are three reasons:

  1. The sample of cases aged 70+ years old was 37.6% in Italy vs 11.9% in China. We know elderly patients have a higher mortality rate overall, hence skewing the CFR in Italy.
    1. A separate report (WHO-China Joint) found that in 80+ year old patients, the CFR was 21.9% and 20.2% in China and Italy, respectively (reassuring that comparable age groups show similar CFR)
  2. The definition of COVID-19 related death in Italy is death with laboratory evidence of SARS-CoV-2 (eg positive test result) independent of pre-existing disease that may have caused death.
    1. The older patient population in Italy has a high burden of co-existing chronic disease (2.7 chronic conditions, on average, per individual) many of which place patients at high risk for developing severe disease.
  3. Testing rates decreased in Italy between February and March
    1. In March, testing became limited to patients requiring hospitalization (representing only the sickest patients more likely to die), thereby reducing the denominator of the CFR (because fewer people, particularly the less severe cases) were being tested
    2. In February, the CFR was 3.1% (when testing was more widely performed).

For the above three reasons, we have hope that the reported case fatality rate is indeed an overestimation. As of now in the USA, the CFR appears closer to China’s reported rate.

The main take-home from this data is that we need to continue shelter in place, to not only protect our high risk and elderly population, but also to not overwhelm our hospitals so that if young and healthy people need to go to the hospital, resources will still be available for them.

Thank you to everybody for following shelter in place and social distancing. Stay safe out there.

Case-Fatality Rate and Characteristics of Patients Dying in Relation to COVID-19 in Italy. JAMA. Published online  March 23, 2020. doi:10.1001/jama.2020.4683


[jetpack_subscription_form title=”Stay Updated” subscribe_text=”Get e-mail updates” subscribe_button=”Stay Updated” show_subscribers_total=”1″]

4 thoughts on “Case Fatality Rate in Italy – Reasons for Overestimation”

  1. Dr. Kaveh, thanks for your great work in covid-19 updates. Lately I’m hearing much about “herd immunization” and posts about having controlled transmission (“vertical spread”). Would you please comment your thoughts on this subject? Thank you.

    1. Hello Edgar,

      Great question. While we strongly suspect that humans are unlikely to become infected twice by the novel coronavirus, we are not yet 100% sure. If our suspicions hold true, this has significant implications, particularly for how the world may “return to work”. This is probably a great idea for an upcoming post 🙂 but in short, if we are able to identify individuals in the population that have acquired secondary immunity (meaning they cannot be affected again), they may be able to reenter the workforce sooner and help our economy pick up sooner (without medical harm).

      “Herd immunization” as a strategy to overcome this crisis faster is not a pleasant scenario to consider (unless we are willing to accept significant loss of life). Assuming the novel coronavirus does not mutate in a way that may re-infect us in the future, it is possible that over longer time periods our species may gain significant immunity against the virus. Though this is also a challenging scenario to consider given the high morbidity and mortality associated with infection (which we would have to undergo prior to full “herd immunization”).

      “Vertical spread” with regards to maternal–fetal transmission, fortunately appears to be rare. We do not yet have a significant amount of data on this topic, but it represents an unlikely source of significant viral spread.

      Does that clarify? Thanks again for a great question.

      1. Thanks for this great clarification, Dr.Kaveh. Sorry about my second question, it was a mistype, I meant “vertical isolation”, where we isolate only the elders and let only young population go out from shelter for work taking some caution like 6-feet distance, no touching, washing hands and so on. Operationally this sounds complex IMO.

        1. This will depend on several factors, including the availability of treatments and effectiveness of shelter in place in the initial weeks. More to come soon, hopefully!

Leave a Reply