GMOA on relative rate of Covid-19 case growth in Sri Lanka: mistaken on both data and trajectory

Fact Check

In its statement, the Government Medical Officers’ Association (GMOA) makes two claims: (i) that the number of Covid-19 cases detected in Sri Lanka in the first seven days since the first local case was detected is higher than that in Italy; and (ii) that this implies a trajectory in which the number of cases will peak during the New Year season, in the absence of aggressive containment measures.

In evidence of its first claim – that the trajectory of confirmed local cases of Covid-19 in Sri Lanka was higher than Italy in the first seven days – the GMOA presented a graph that has been recreated in Exhibit 1.

In comparing Sri Lanka’s trajectory with that of Italy, the GMOA makes reference to the first ‘local’ case detected. This is a standard distinction made in counting cases, where those that contracted the virus outside the country and brought it in are distinguished as ‘imported’ cases, whereas ‘local’ cases are those that contracted it in the country.

FactCheck evaluated the GMOA’s first claim using data from the Johns Hopkins Coronavirus Resource Center, which tracks the number of detected cases of Covid-19 detected for all countries on a daily basis. Additionally, FactCheck also consulted local and international news reports to obtain more information on the Covid-19 cases detected in both Sri Lanka and Italy.

Exhibit 2 shows the daily trajectory in the number of local cases for Sri Lanka and Italy. It excludes the first reported imported cases in both countries, i.e., Case 1 detected on 27 January in Sri Lanka; and Cases 1-3 in Italy detected during the period 31 January – 7 February, which were all imported cases. Italy’s first locally contracted case was detected only on 21 February.

Exhibit 2 shows that on the day the GMOA made its statement and presently as well, Sri Lanka is on a lower trajectory compared to Italy in terms of the number of cases detected. The mistake in the GMOA data arises, possibly, from either getting the date of the first local case wrong, or incorrectly using Italy’s imported Case 3 as the first local case. Thereby, the GMOA arrives at an incorrect conclusion about relative growth of case detection in Sri Lanka, in relation to Italy in the first seven days. Therefore, we classify the GMOA’s first claim as false.

The second part of the GMOA’s statement also suggests that Sri Lanka’s spread will peak during the Sinhala and Tamil New Year period in April, if mitigation measures were not taken. This raises the question of not just the initial trajectory but also the future trajectory of the virus spread in Sri Lanka.

Therefore, the second part of this fact check will aim to shed further light on (i) the growth of Sri Lanka’s reported cases relative to other countries; and (ii) when Sri Lanka’s detected cases could peak, based on the best available estimates and models of disease spread that are accessible at present. The modelling is in early stages and provided as the best available estimates in a context of many unknowns, and not as a factual claim.

Exhibit 3 uses the data obtained from the Johns Hopkins Coronavirus Resource Center to compare the trajectory of Sri Lanka’s increase in detected cases, from the tenth case onwards, against all other countries that have ten or more cases.  It shows Sri Lanka to be on a trajectory that is about midway between the most extreme situations (such as Italy) and the least extreme situations (such as Taiwan).

There is a constant flow of new information with regard to Covid-19, and it is undoubtedly difficult, as yet, to speak with certainty on the future spread of the virus. Epidemiological modelling of the spread depends critically on the assumption made about the rate of spread – that is, how many new people will contract the virus through an infected individual during their infection. Let’s call this number r. For example, if r = 3, this means that an infected person will infect 3 more people during the course of their infection. The value of r is significantly determined by the mitigation strategies that a country puts in place to reduce person-to-person contact. The more aggressive the strategy, the lower the contact between infected and healthy individuals, and thus the lower the value of r.

Exhibit 4 shows draft initial results from the standard epidemiological models that Verité Research uses to simulate infection rates for Sri Lanka, based on active cases as at 28 March 2020. It shows how the disease spread could play out for different levels of r under different mitigation strategies. If mitigation measures are too mild, and r increases beyond 2, then the health care system may not be able to cope with the rate of growth in relation to even the small percentage of infected cases that would need hospital care. The strongest mitigation strategy, modelled as providing a value of r=1.5, show Sri Lanka overcoming the virus spread after 257 days or 8.5 months since 28 March 2020, with the least pressure on the healthcare system. As Exhibit 4 shows, even with no mitigation strategy in place (modelled by the scenario where r = 3.5, which is similar to the rate of spread in Italy before mitigation strategies were enforced), the peak level of infection would occur in early June, well beyond the Sinhala and Tamil New Year season in mid-April. Therefore, the second claim by the GMOA is also demonstrably incorrect.

Note: A suppression strategy for managing the spread would imply a value of r<1. This means that it would take more than one infected person to infect a healthy person. For example, if r = 0.5, it would take two infected persons to infect a new person.  This case has not been modelled because it would result in Sri Lanka being back at square-one after the suppression measures are lifted and the virus begins to spread again before a vaccine is found. The lock-down measures needed for suppression measures may also not be sustainable, due to the economic survival consequences on human lives, which will need to be evaluated alongside the human cost of the spread of the virus.

*FactCheck’s verdict is based on the most recent information that is publicly accessible. As with every fact check, if new information becomes available, FactCheck will revisit the assessment.