Predictions for COVID-19 outbreak in India using Epidemiological models

Rajesh Ranjan

This exploratory evaluation shows that the transmission rate of COVID-19 in India is comparable to that in Washington state in the US. The curves describing the initial phase of the outbreak for both the locations are very close. As the beginning of the outbreak in Washington was about 9 days prior to that in India, the current data from the former can be used to make informative predictions for India. Along the similar argument, we show that the transmission rate and basic reproduction number R 0 for India are in the expected range. This is contrary to several reports questioning the testing strategies adopted by India quoting a low transmission rate and hence the smaller basic reproduction number. Two epidemiological models -a simple exponential model and an SIR model, are used respectively to forecast short and long term outcomes. These models assume all the seed cases to be symptomatic, which may underestimate the actual numbers due to an uncertain number of asymptomatic individuals. With this limitation, the exponential model based on data till March 23 nicely predicts the values till today (March 30). The SIR model based on data till March 30 indicates that India will enter equilibrium by the end of May with an estimated total number of infected cases to be approximately 13,000. It is estimated that the impact of social distancing will be visible after April 8 following which we may see a significant reduction in the reported infections. However, India is on high risk to enter into community transmission due to reported violation of quarantine norms by individuals as well as other social and demographic issues. The predictions made using the current epidemiological models in the current work will be invalid if such an event occurs. Finally, the model is as good as the underlying data. Because of real time change in data daily, the predictions will accordingly change. Hence, the results from this paper should be used only for qualitative understanding and reasonable estimate of the nature of outbreak, but are not advisable for any decision making or policy change.