Active cases in the ongoing second COVID-19 wave in India may peak at 38-48 lakh between May 14-18 and daily new infections could hit a high of 4.4 lakh from May 4-8, according to a mathematical model by IIT scientists who have revised their projections upwards.

India on Monday saw a single-day rise of 3,52,991 (3.52 lakh) COVID-19 infections and 2,812 fatalities with 28,13,658 (28.13 lakh) active cases.

The scientists from the Indian Institute of Technology in Kanpur and Hyderabad applied the Susceptible, Undetected, Tested (positive), and Removed Approach'' (SUTRA) model to predict that active cases would go up further by over 10 lakh by mid-May.

The latest projection tweaks the time frame as well as the numbers.

Last week, the researchers predicted the pandemic may peak between May 11-15 with 33-35 lakh total active cases and decline steeply by the end of May.

Earlier this month, their modelling approach projected that active infections in the country would peak by April 15, which didn't come true.

“This time, I have also computed minimum and maximum for predicted values and posted it. I am reasonably confident that the actual values will be within the min and max values mentioned,” Maninder Agrawal, professor at the Department of Computer Science and Engineering, IIT-Kanpur, told PTI.

On Sunday, Agrawal shared the new peak values for active and new COVID cases in a Twitter thread.

“Peak timing: May 14-18 for active infections and May 4-8 for new infections. Peak value: 38-48 lakhs for active infections and 3.4 to 4.4 lakhs for new infections,” he said.

He also noted that it was not clear what the final values would be.

“I have now computed a range of values for peak value and timing and the final numbers should be within this range,” Agrawal said in another tweet.

Explaining the changing projections of the model over time, Agrawal said, “The primary reason is that there is a continuous slow drift in parameter values of current phase for India. This makes the guess of the right values difficult. That is causing a slow change in predictions.” The scientists in the as yet unpublished study said there are several novel features in the SUTRA model.

Whereas previous papers divided the patient population into asymptomatic and Infected, the new model also accounts for the fact that some fraction of asymptomatic patients could also be detected due to contact tracing and other such protocols.

The IIT Kanpur professor noted that the SUTRA model uses three main parameters to predict the course of the pandemic.

The first is called beta, or contact rate, which measures how many people an infected person infects per day. It is related to the R-nought value, which is the number of people an infected person spreads the virus to over the course of their infection, Agrawal explained.

The other two parameters are ‘reach'', which is a measure of the exposure level of the population to the pandemic, and ‘epsilon'' which is the ratio of detected and undetected cases.

Independent calculations by Gautam Menon and his team at Ashoka University in Haryana had predicted that the peak of the ongoing wave of infections could be between mid-April and mid-May.

Menon also cautioned that such projections of COVID-19 cases should really be trusted only in the short term.

Any excessively precise prediction, of a peak within just a five-day window would ignore the many uncertainties associated with the inputs to any such calculation, Menon, who was not involved in the modelling, had told PTI earlier.