Predictions for COVID-19 outbreak in India using Epidemiological models

Rajesh Ranjan

As the current model is fully dependent on data, it is imperative to comment on the nature of this data. Different countries have different strategies for conducting COVID-19 diagnostic tests. In India, testing has largely been limited to individuals travelling from high-risk countries and their immediate contacts, as well as selected Pneumonia patients and symptomatic healthcare workers. As of March 30, India has tested 42,788 samples. Indian medical authorities have justified this strategy by testing randomly collected samples. Sahasranaman and Kumar [10] have compared basic reproduction number R 0 from India and the world to analyze this strategy. R 0 is the transmission rate given that the population has no immunity from past exposures or vaccination, nor any deliberate intervention in disease transmission. The number of infections grow and spread in the population if R 0 > 1. They have found that R 0 from India (R 0 ≈ 0.43) is much smaller than the rest of the world (1.5 < R 0 < 2.5), and the numbers reported in India may not be reflective of the actual number of cases. However, by considering a longer date range from March 10 to March 30, we show that the value of R 0 from India is comparable to infection rates reported elsewhere. Further, we find that the growth of infections in India is comparable to that in California and Washington. In any case, current models do not depend on the testing strategy provided the same protocol is used throughout the time-period. In other words, the ratio of actual to reported number of cases will remain the same at any given time. On the other hand, the uncertainty due to exclusion of asymptomatic cases can be a major limitation in predictions with these models.