Izvestia: A rise with a turn: Mathematicians come up with a new forecast for Covid cases
By mid-November, the number of new cases of coronavirus disease per day may range from 26,500, according to the optimistic scenario, up to 35,000, according to the pessimistic one, mathematicians at St Petersburg University have calculated. They made an updated forecast using a newly optimised model.
According to their calculations, by mid-November the number of COVID-19 patients will be within the range of 920,000 up to 1 million people. Over recent weeks, Covid cases have been on the rise in Russia. On 11 October, the number of new cases per day reached 29,409. This is a record since the beginning of 2021.
Stepping up the ladder
In mid-September, the Covid cases started to rise again. On 11 October, the number of new cases per day reached 29,409.
’If the number of cases continues to rise and by early November the increase in percentage reaches 0.41%, the number of actively ill people (the total number of cases minus the total number of recovered and deaths. By Izvestia) is calculated to reach 1 million people in the first week of November’, said Viktor Zakharov, Head of the Department of Mathematical Modelling of Energetic Systems at St Petersburg University and Head of the Centre for Intellectual Logistics at St Petersburg University. ‘This is expected to coincide with the pandemic record number of daily new cases of 35,000.’
However, if the percentage increase remains at the current level, by 10 November the number of cases can only rise up to 33,000 cases per day, he said. The number of active cases, as he put it, will approach 1 million and start to decrease.
‘Following the optimistic scenario, the number of daily new cases will drop to 26,500 by 10 November,’ added Viktor Zakharov. ’The number of active cases will reach 920,000 and start to decline.’
Knowing where to stop
Vladimir Mazalov, Director of the Institute of Applied Mathematical Research at the Karelian Research Centre of the Russian Academy of Sciences, agreed with the pessimistic scenario developed by the colleagues. He believes that the number of new Covid cases will grow. Roman Zinovkin, Leading Research Associate at the Laboratory of Molecular Biology at Lomonosov Moscow University, is also inclined to follow the disappointing forecast.
’This is a very realistic forecast, just look at the incidence graphs,’ the expert said. ’The reasons behind are rather trivial. First, the beginning of the school year and the spread of the coronavirus by children. Second, the Delta strain is spreading among those who have been vaccinated, although they definitely much more easily come through the disease than those who have not been vaccinated. Vaccinated people are much less at risk. Thirdly, people are tired of restrictions, people are less vigilant in wearing masks and observing social distance. If everyone had been wearing masks, we would not witness such a rise in Covid cases.
The most optimistic scenario is possible if more serious restrictions throughout the country are introduced. If everything remains the way it is, the pessimistic version is more probable, he said.
35,000 new cases per day in November is a realistic scenario and the model is quite true to what might happen, agreed Artur Karimov, Senior Research Associate at Youth Research Institute at St Petersburg Electrotechnical University ‘LETI’.
‘Another reason behind the increase in the number of cases is seasonality and consequences of increased mobility in summer,’ the expert said. The tourists who returned to Russia from abroad are actively spreading new strains. Additionally, combination of influenza and Covid in autumn is particularly dangerous. I would urge everyone to be vaccinated against both infections. A full decline in the number of cases, in my opinion, will begin no earlier than 15 December.’
Ensuring accuracy
The researchers have long been using the mathematical model. However, they have just recently optimised it, which ensures more accuracy in calculations.
‘We have introduced a new parameter in our model, i.e. the characteristic of dynamic balance,’ explained Viktor Zakharov. ’It links the total number of infected at a particular moment in the past and the total number of people who recovered or died by the selected day.’
The use of the principle of dynamic balance can significantly increase the accuracy of the forecast of the future dynamics of active cases on the horizon of four to five weeks, said the expert.
Along with the balance model, the new approach to forecasting is based on the model for estimating the total number of infected people that was proposed by St Petersburg University. It is based on an iterative approach, i.e. the data to forecast for two to three weeks is updated in real time.
‘Now we have a different model. In other words, it is different from the so-called SIR (an abbreviation for Susceptible, Infected, Recovered — susceptible to a disease, infected and recovered. By Izvestia), where the dynamics of the pandemic is based on the frequency of meetings of healthy and infected people’, said Vladimir Mazalov. ’The problem of using this model is that there are many parameters to be evaluated. At first they are unknown, when they accumulate, the model operates more accurately, yet still with errors. The model used by Viktor Zakharov is simpler. The data of their modelling are regularly confirmed by real data.
Earlier, the forecast that was based on the new model came true. At the end of September, Viktor Zakharov’s team predicted a rise in the number of cases. By mid-October, the scientists said, the total number of patients will grow up to 700,000, while the number of new cases will be 27,000 per day. On 11 October, the number of active cases amounted to 713,823 people. Earlier, during the second peak in the incidence of coronavirus, the maximum number of active cases was recorded on 8 January 2021 and amounted to 563,754 people.