By Dushan Wadduwage, Savithru Jayasinghe, Udeni De Silva Perera
Today, COVID-19 has become a regular phenomenon that affects all aspects of day-to-day life in Sri Lanka. While we had less than 50 new cases and rarely any deaths the same time last year, we now report thousands of daily new cases and close to 50 deaths every day. Considering that it took more than 7 months from the first reported death on March 28th last year to reach a total of 50 COVID deaths, reaching a daily average death-toll of close to 50 would have been unthinkable the same time last year. Given this rapidly changing landscape, a prudent question to ask would be whether this is as bad as it would get, or whether the worst is yet to come?
Contrary to general wisdom, this question is easy to answer from the basic principles of epidemiology. Things could get worse. As long as the population largely remains unvaccinated, we are still at what epidemiologists call the early stages of the pandemic. Even if we assume that the actual number of COVID cases is ten times higher than what has been reported, the percentage of the population who has gained immunity through infection is less than 10%. This, together with the fact that at the time of writing less than 10% of the population is fully vaccinated, means that approximately 8 out of 10 people remain susceptible to the disease. Our estimates suggest that, during the month of April, the effective reproduction number (R) of COVID-19 in Sri Lanka was well above the critical value of 1, and during the peak super-spreader period,was close to 1.7! Through mathematical modeling, we can unambiguously predict that similar R values would initiate another covid wave. Things could definitely get worse again!
So how do we prevent another wave? First and foremost, let’s quantitatively understand where we are in the third wave. To quantitatively model a pandemic such as COVID-19, scientists often make use of two major data streams: daily reported infections and daily reported deaths. Given the clinical progression of the disease, these two data streams can be used to fit mathematical models to predict future scenarios. However, when the healthcare system of a country is overwhelmed by a sudden explosion of cases, infections are often under-reported, and the accuracy of these data streams come into question. Due to the small number of cases during early times of COVID-19 in Sri Lanka, i.e. the first and the second waves, we can assume that infections were correctly reported. This allows us to calculate the infection fatality ratio (IFR) of the first and second waves to be around 0.55%. However, there could be a few different possibilities with regards to the much larger third wave, which started in early April of 2021.
Given we have not yet faced as dire a state as our neighbour India, we can safely assume that the deaths are by and large correctly reported in the third wave. In that case, If there is no under-reporting of cases and the IFR of the third wave is the same as the first and the second, i.e. 0.55%, then the mathematical model predicts a lower number of deaths than the actual deaths (Figure 1). This contradiction is a classic case of under-reported infections. By calibrating the model to match forecasted deaths with actual deaths (Figure 2), we can in fact estimate that the actual number of cases to be at least 3 times higher than the reported cases. For instance, although the reported number of cases was 2877 on June 1st, the actual number of cases could have been over 11,000!
The second scenario is that the IFR for the third wave could be different from that of the first and second waves whilst the reported number of infections is mostly accurate. In such a situation, the model can be tuned to agree with both the reported number of cases as well as deaths, if the IFR is increased up to 1.5% (Figure 3). This is unlikely, but possible for two reasons. First, the third wave could have been driven by a more aggressive and fatal variant which may result in a higher IFR. Second, the level of patient care may have deteriorated during the third wave resulting in a higher number of fatalities (i.e. higher IFR). Either way, health officials need to investigate these possibilities with additional data in order to zero-in on the real scenario. Such an effort will provide valuable information to plan precise solutions to the current crisis in the short term.
In part 2, we will quantitatively analyse the exit strategies for the current situation.
