Probabilistic approximation of effective reproduction number of COVID-19 using daily death statistics.

The effective reproduction number () which signifies the number of secondary cases infected by one infectious individual, is an important measure of the spread of an infectious disease. Due to the dynamics of COVID-19 where many infected people are not showing symptoms or showing mild symptoms, and where different countries are employing different testing strategies, it is quite difficult to calculate the , while the pandemic is still widespread. This paper presents a probabilistic methodology to evaluate the effective reproduction number by considering only the daily death statistics of a given country. The methodology utilizes a linearly constrained Quadratic Programming scheme to estimate the daily new infection cases from the daily death statistics, based on the probability distribution of delays associated with symptom onset and to reporting a death. The proposed methodology is validated in-silico by simulating an infectious disease through a Susceptible-Infectious-Recovered (SIR) model. The results suggest that with a reasonable estimate of distribution of delay to death from the onset of symptoms, the model can provide accurate estimates of . The proposed method is then used to estimate the R values for two countries.