Two-thirds of Covid-19 cases may be undiagnosed: modelling estimate based on real-world data

GlobalData Healthcare 20 April 2020 (Last Updated April 20th, 2020 14:32)

Two-thirds of Covid-19 cases may be undiagnosed: modelling estimate based on real-world data

The coronavirus disease 2019 (Covid-19) has now spread across 187 countries with approximately two million confirmed cases worldwide. The US is the most heavily affected country in the world, representing 25% of confirmed cases worldwide. Based on modelling assumptions using available data to date, GlobalData estimates that 50–80% of the Covid-19 cases are mild and asymptomatic and thus not likely to be diagnosed. It is paramount that undiagnosed cases of Covid-19 are mapped quickly to accurately predict the number of asymptomatic contagious cases. Having an accurate estimate of the infected numbers will also be important in helping governments plan for the re-opening of society.

The true number of untested/undiagnosed cases is creating a lot of debate as well as confusion. The limited data from China in the early period of the outbreak showed that more than 80% of cases could be asymptomatic and not likely to be diagnosed. This suggests that there could be eight million cases worldwide that have not been diagnosed. The World Health Organization (WHO) had also corroborated during the early period of the outbreak that 80% of infections are mild or asymptomatic, while 15% are severe and 5% is critical. These assumptions were made with very thin data, including only initial outbreak data from Hubei province in China and early outbreak data in Italy.

The newest data from Austria, Iceland, and New York show variations in undiagnosed estimates of Covid-19. In Austria, a random sampling survey reported that 0.33% of Austrians had active Covid-19, which suggested that 50% of cases are undiagnosed. Iceland has a policy of widespread testing that did not limit testing to those with symptoms. Data from Iceland reported that 50% of cases were asymptomatic carriers. Data from a study in New York of all women giving birth at a hospital showed 88% of the pregnant women who tested positive for the novel coronavirus were asymptomatic carriers. The Centers for Disease Control (CDC) report that 25% of cases were asymptomatic suggests that the proportion of novel coronavirus asymptomatic carriers is in the 25–88% range.

GlobalData epidemiologists forecast the total number of cases (diagnosed and undiagnosed) in Austria, Iceland, China, and New York using these real-world data. Cumulative confirmed cases on 15 April were used as a baseline in this model, shown in Figure 1. New York had 640,000 cases of Covid-19 and two-thirds of those cases are undiagnosed. Approximately 250,000 cases in China are still undiagnosed, whereas undiagnosed cases represent 50% of total cases in Austria and Iceland.

Finally, if we extrapolate the total infection rate (undiagnosed and diagnosed) from these countries to the rest of the world, no country is close to herd immunity, where at least 60% of the population is exposed to the virus and can build up a combined immunity. This means countries need to be vigilant of second and third waves of infection, or even seasonal outbreaks that could arise out of both uncontained local pockets and new introductions from travellers once lock-down measures are relaxed. It may be prudent to implement slow release of lock-down measures so governments can better monitor and address potential secondary outbreaks.

Figure 1:  Estimate of Combined Undiagnosed and Diagnosed Cases of Covid-19, N, 15 April 2020