Updated on March 27th 2020, written by Dr David Owens.
To understand the risk of catching COVID-19 we need to understand the epidemic.
Understanding epidemics is maths and data. To try to simplify this complex process I will describe the epidemic curves (below) and continue the analogy of the epidemic as a fire.
Think of this epidemic as being like a fire which started to rage in Wuhan. We didn’t have methods to put out the fire (drugs to treat the illness) so we needed to focus on containment. The fire will need to burn out by itself. Sometimes nature helps us, although in the case of the coronavirus it is sunshine and heat that help to put out the fire rather than rain. This is how the public health measures work. We built a wall around the fire to contain it. As often happens in this situation some embers escaped. Some landed on the ground and burnt out but others started smaller, little fires. Because the alarm had sounded everybody was now aware and focused on the new outbreaks and it is usually easier to contain small fires than big ones. There is a danger that an ember lands in a particularly dry and vulnerable place and triggers another larger fire (or series of small fires). We can consider that this has happened on the cruise ship, in Iran, Korea and Italy.
How well the fires are contained in new locations depends on a number of factors, including population immunity, environmental and political factors which influence the ability of the medical firemen to do their job of containment and the capacity of the health system to manage the resulting complications of the illness. Infectious disease (fire in this analogy) always does the greatest damage in poor communities with low capacity health systems. The Director General of the WHO made increasingly vocal appeals asking countries to prepare. After relatively fragmented response the world reacted in dramatic fashion following the declaration of a pandemic. Widespread social distancing policies including closure of borders, restriction of international travel and quarantine has been introduced throughout the world.
I have explained in another article the reasons that we close schools. We do not close schools because of the impact of this illness on our children. Closing schools is like spraying water on the ground. It makes it less likely that an ember will take hold and start a new fire. Public health controls work on the principle that we minimise spread, boost population immunity and contain all the little fires by quarantining close contacts of individuals who are infected.
All epidemics of infectious disease can be described in terms of a basic reproduction number R0 for each illness. This number is a measure of on average how many people are infected by an individual with the specific disease. If this number is <1 the epidemic will die (this means that on average each infected person infects slightly less than one other person). If this number is >1 the epidemic is likely to grow. Infections with higher R0 values are more likely to spread. Population measures such as isolation of infected individuals, masks, hand washing, school closures and social distancing all work by reducing the exposure risk with the intention of reducing the effective transmission below 1 but regardless as much as possible in order to give the best possible chance for the epidemic to die. In China, the public health measures reduced the early Ro measures from 3.5 to 0.38. They were highly effective. How effective the measures will be in Europe by comparison will be key in controlling the epidemic.
Below is the epidemic curve for the COVID-19 in China. This is a measure of the new cases being diagnosed each day.
Figure 1 - Daily number of newly confirmed cases reported in Mainland China since January 10, 2020 (including cases based on clinical diagnosis from Hubei Province since February 12, 2020). The sharp jump in the graph was when reporting criteria in Hubei was changed from diagnosis based upon test results to clinical diagnosis. This made sense due to resource limitations of testing kits at that time.
Figure 2 is the latest data from the epidemic curve for Guangdong1.
Figure 2 - Daily number of confirmed cases reported in Guangdong Province since Janaury 10, 2020
This is why the authorities in China were confident that the epidemic had been brought under control. The public health measures and the aggressive containment response in China was highly effective. This suggests that the main fire may have burned out. Of course there is no room for complacency, fires can start up again or spread somewhere else. This has happened in different parts of the world. The WHO produces a consolidated epidemic curve for cases outside of China2. We can consider this to be a measure of the combined intensity of all the different fires in different parts of the world. We can see that this is increasing. It is important to appreciate that as the fire in China has been controlled fires are now burning all over the world. The increasing incidence of new cases in more locations can be thought of as like increasing embers, each with the potential to set off a larger fire. The containment and control of these new cases will be critical in minimising the total epidemic size.
Figure 3 - Epidemic curve of confirmed COVID-19, by date of report and WHO region through 26 March 2020.
In order to get a broad perspective of the evolving epidemic we must consider the epidemic curves of all of the clusters independently. This will tell us where the fires are burning and importantly where they are increasing and reducing in intensity. Below is the epidemic curve of Korea. This explains why the DG of the WHO expressed some optimism about containment efforts in Korea.
Figure 4 - Daily number of newly confirmed cases reported in Korea since January 10, 2020. An analysis of the breakdown of the cluster of cases in Korea can be found here3.
Figure 5 - Daily number of newly confirmed cases reported in Japan since January 10, 2020.
Figure 6 - Daily number of newly confirmed cases reported in Italy since January 10, 2020. An analysis of the breakdown of the cluster of cases in Italy can be found here4.
Italy introduced intensive social distancing policies on March 9th and 11th 2020. Following the Italian cluster and comparing with Hubei will give useful information about the behaviour of the fire under different containment strategies. See if the curve starts to flatten which would be a hopeful sign that similar policies may work in other European countries. An epidemiological model of the Italian public health regulations has been reported in the Lancet as showing some early and positive signs in comparison to non-intervention5. This suggested some benefit within 3-4 days of the regulations. Obviously it will be important if this continues to have impact.
Figure 7 - Daily number of newly confirmed cases reported in Iran since January 10, 2020.
Figure 8 - Daily number of newly confirmed cases reported in France since January 10, 2020.
Figure 9 - Daily number of newly confirmed cases reported in Germany since January 10, 2020. The newly confirmed cases on March 9 2020 and March 18 2020 included data from 08:00 to 15:00 and from 15:00 to 00:00 local time in Germany respectively.
Figure 10 - Daily number of newly confirmed cases reported in Spain since January 10, 2020.
Try to look at the change in cases rather than cumulative cases. Many of the newspapers focus on the cumulative data. The number of cases and the number of deaths will get bigger every day. It is the rate of change which tells us whether the fire is getting bigger or smaller. Epidemic curves are useful to get a sense of where the fires are burning and most importantly whether that intensity is increasing or reducing as this will determine risk. It is important to realise that even within countries, we are not dealing with one big fire, but lots of little clusters which may be very distinct in the risk that they pose. Remember too that death is a lagging indicator. People must first get sick, they are then usually diagnosed and either recover or in some cases die. In the first series the mortality was 10-12 days post diagnosis. This may reasonably expected to be delayed with greater awareness and understanding of treatments so look for the mortality curve to lag the new cases by a number of weeks.
It is important to realize that the increase in cases are often now from contacts. These are people who have been screened and tested rather than those in the early cluster who went to hospital because they felt ill. Information about the milder cases will be very important. As explained previously, the more mild cases the less severe this illness is likely to be on a case by case basis.
A pandemic is simply the name given when fires are burning on different continents. The name doesn’t change anything. The impact in population terms will depend on how bad the fires are and how well they can be contained.
Ultimately the impact of this disease on the population will be driven by the evolution of the epidemic. Excessive focus on the disease has the potential to distort real risk and significantly increase anxiety for the reasons previously discussed.
Information and education are key factors in the management of both anxiety and infectious disease. We have discussed the importance of having an anchor against which to measure risk. We have chosen influenza as a known entity against which to compare risks. COVID-19 is not a type of influenza and we are certainly not down playing its importance. Currently this disease seems to be more severe. Until we know the number of mild cases we cannot be certain about the relative severity on a case by case basis. Influenza kills 650,000 people globally from respiratory disease. Although the numbers for COVID-19 are currently significantly smaller they are accelerating rapidly. Whether COVID-19 has a greater or lesser disease burden than influenza will be determined by the ultimate size of the epidemic. This will depend in large part on the success of the public health interventions and comparison will only be possible in retrospect. Look at the changing numbers rather than cumulative COVID-19 numbers in isolation in order to get a sense of changing risk.
1. (2020). Retrieved 5 February 2020, from https://www.chp.gov.hk/files/pdf/statistics_of_the_cases_novel_coronavirus_infection_en.pdf
2. Coronavirus disease 2019 (COVID-19) Situation Report – 66. (2020, March 26). Retrieved from https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200326-sitrep-66-covid-19.pdf?sfvrsn=81b94e61_2
3. Centre for Health Protection. “Latest Situation of COVID-19 in Korea and Revised Reporting Criteria for Suspected Cases of COVID-19.” Latest Situation of COVID-19 in Korea and Revised Reporting Criteria for Suspected Cases of COVID-19, Feb. 24AD, www.chp.gov.hk/files/pdf/letters_to_doctors_20200224.pdf
4. Centre for Health Protection. “Latest Overseas Situation of COVID-19 and Revised Reporting Criteria for Suspected Cases of COVID-19.” Latest Overseas Situation of COVID-19 and Revised Reporting Criteria for Suspected Cases of COVID-19, 28 Feb. 2020, www.chp.gov.hk/files/pdf/letters_to_doctors_20200228.pdf
5. Saglietto, A., D'Ascenzo, F., Zoccai, G. B., & De Ferrari, G. M. (2020, March 24). COVID-19 in Europe: the Italian lesson - The Lancet. Retrieved from https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)30690-5/fulltext