Improved SIR - Covid-19 Prediction Model

SIR (Susceptible-Infected-Removed) model to estimate the peak while a data-driven approach based on past outbreaks is used to predict the decline of the epidemic.

Data Insights

Date: 25-05-2021

Delhi has alarmingly high average case fatality (CFR) rate of 9% compared to all other states. National CFR is less than 2%. Kerala has CFR of about 0.5%, while Maharashtra, UP and Punjab all have CFR close to 3.5%.

Date: 29-05-2021

Most of the states in India except a few in the Northeast region have already peaked. However, they all may not open at the same time. This is due to differential test positivity rates (TPRs) among the states. TPR is defined as the number of positive cases per test conducted and therefore is a measure of spread of the pandemic. Higher the TPR, higher the spread, and vice versa. The figure shows 7 states who are in decline. Among these, UP and Delhi have TPR less than 5% and are likely to open soon, of course with necessary cautions. Karnataka and Kerala have high TPR (>20%), so it may take some before the interventions are eased there. WHO recommends TPR to be less than 5% for at least two weeks before interventions are eased. TPR for other states can be accessed from respective pages on the website.

Date: 18-06-2021

Here is what our SIR model suggests about the possible third wave, assuming that India has not reached herd immunity already. We obtain the epidemic parameters from the second wave and assume that the complete unlocking begins on 15th July 2021. We simulate three scenarios: 1) Normal i.e. post unlocking the situation is similar to that in early 2021, 2) Normal but with virus mutations i.e. more infectious variant (such as delta plus) comes into play, 3) Stricter interventions i.e. higher compliance than that in early 2021 is maintained. Scenarios 2 and 3 are simulated by respectively increasing and decreasing the contact rate by 5% than the 'normal' level. The positive impact of vaccinations is not considered explicitly but can be implied from scenario 3. Premature or very late unlocking ( than the current date- 15 July) could alter the situation drastically.

Few observations:
(a) Normal (and most likely) scenario suggests the third wave peak in October. With same parameters as the second wave, a lower peak than the second wave is obtained. Effective vaccination can further help in lowering this peak.
(b) If the virus mutates and stricter interventions are not followed (scenario 2), the peak could be higher than the second one and come early (September).
(c) If people follow strict social distancing measures (scenario 3), the peak could be delayed until late October, and also the peak height will be much lower than the second wave (1.6 lakhs). Further, a relatively long time could help in vaccinating a critical mass for herd immunity, and therefore the third wave can be avoided altogether.
In a nutshell, do not lower the guard. If we all are careful and get vaccinated at the first opportunity, we can escape the third wave. Stay safe and maintain social distancing!