Levitt Model for Covid-19: No room for complacency

A disease prediction model proposed by 2013 Nobel laureate Michael Levitt that relies on a simple mathematical formula was recently applied to India and various states by Prof. Bhaskaran Raman, Indian Institute of Technology (IIT), Bombay. Several researchers have already highlighted the limitations of this model such as assuming a linear extrapolation of the growth rate and not factoring the time-dependent nature of the reproduction number (Rt). We are in agreement with them that this model is perhaps more suited for “explaining” past data to some extent and that is what we intend to use it for as well.

What has emerged so far from Prof. Bhaskaran Raman’s analysis using the Levitt model is that COVID related deaths in states such as Delhi and Tamil Nadu are showing a downward trend whereas on the contrary Karnataka is one of the few states seeing an upward trend during the same time period. Now, instead of using this model to predict when the pandemic will perhaps end, it would be interesting to use this model to analyze and dissect what has happened so far.

It should be noted that Delhi too had an upward trend in COVID deaths (as currently observed in Karnataka) especially during the time period between 40 to 90 days from the first reported COVID death.

However, there was an inflection point at approximately 90 days and subsequently there has been a downward trend in COVID deaths as highlighted by Prof. Bhaskaran Raman.

A similar trend is observed in Tamil Nadu as well. An upward trend in COVID deaths is observed especially during the time period between 40 to 84 days from the first reported COVID death.  Post that, there is an inflection point at approximately 84 days and since then there has been a downward trend in COVID deaths, again highlighted by Prof. Bhaskaran Raman.

So, why is Karnataka having an upward trend when other states like Delhi and TN are seeing a downward trend in COVID deaths?  In the case of Karnataka, during the first 90 day period itself we see a very slight upward trend followed by a flattening to eventually a very slight downward trend.  These 90 days coincide with the nationwide lockdown phase till May 3rd 2020 followed by the continued lockdown with a few relaxations till June 8th 2020.  It looks like that within these first 90 days, Karnataka even had a mild peak which was very well managed through rigorous contract tracing, tracking and isolation – also hailed as the “Karnataka model” even by the Prime Minister.

Thus, during the time period between 40 to 90 days (since the first COVID death) when states such as Delhi and TN were seeing a surge, Karnataka on the other hand had successfully mitigated the situation. In retrospect, if the Levitt model would have been applied to Karnataka at that point in time, it would have most likely predicted the end of the pandemic in Karnataka by now!

However, the situation soon turned for the worse. After the first 90 days, the famed “Karnataka model” began to slowly disintegrate and the state started seeing a strong upward trend in COVID deaths.  It is almost as if the “real” pandemic has just begun.

One of the valuable lessons that states like Delhi and TN can learn from Karnataka is that even if they reach a situation of minimal growth rate in daily COVID deaths, there shouldn’t be any complacency with regard to the 5T’s (Tracing, Testing, Tracking, Treatment and Technology).  The Levitt model may perhaps not be able to accurately answer or predict complex scenarios like the one emerging in Karnataka, but it definitely helps in raising relevant questions.

[Disclaimer: Views presented above are those of the author and do not necessarily reflect those of CBPS]

Sridhar Prasad
Research Advisor, CBPS

Leave a Reply

Your email address will not be published. Required fields are marked *