Prediction of daily and cumulative cases for COVID-19 infection based on reproductive number (R0) in Karnataka: a data-driven analytics

Abstract

Abstract To estimate the reproductive number (R0) of the coronavirus in the present scenario and to predict the incidence of daily and probable cumulative cases, by 20 August, 2020 for Karnataka state in India. The model used serial interval with a gamma distribution and applied ‘early R’ to estimate the R0 and ‘projections’ package in R program. This was performed to mimic the probable cumulative epidemic trajectories and predict future daily incidence by fitting the data to existing daily incidence and the estimated R0 by a model based on the assumption that daily incidence follows Poisson distribution. The maximum-likelihood (ML) value of R0 was 2.242 for COVID-19 outbreak, as on June 2020. The median with 95% CI of R0 values was 2.242 (1.50–3.00) estimated by bootstrap resampling method. The expected number of new cases for the next 60 days would progressively increase, and the estimated cumulative cases would reach 27,238 (26,008–28,467) at the end of 60th day in the future. But, if R0 value was doubled the estimated total number of cumulative cases would increase up to 432,411 (400,929–463,893) and if, R0 increase by 50%, the cases would increase up to 86,386 (80,910–91,861). The probable outbreak size and future daily cumulative incidence are largely dependent on the change in R0 values. Hence, it is vital to expedite the hospital provisions, medical facility enhancement work, and number of random tests for COVID-19 at a very rapid pace to prepare the state for exponential growth in next 2 months.