- An online epidemic predictor tool can give an idea how hard covid-19 can hit India.
- The tool is based on SEIR model that experts use to predict epidemics in a region.
- The online tool is made by Gabriel Goh, a machine learning researcher and can be accessed with a web browser.
What is going to be the future of Covid-19 in India? For how long coronavirus will infect people here? How many people are going to fall sick with Covid-19 in India? And how many people may die of Covid-19 here? No one has a sure answer to these questions. Possibly because no one can have an answer to these questions. But in case you are curious there is an online tool created by a machine learning expert that shows the course Covid-19 pandemic may take in India. It is a simple tool, but one based on SEIR model that scientists and public health experts use to predict epidemic in a given region.
The tool has been created by Gabriel Goh, a machine learning researcher.
There is nothing official about the tool but during a time when the information from government experts may not be coming or coming fast enough it can shed some clarity on what Covid-19 can do to India. Depending on what measures the country takes, and what impact these measures have on Covid-19 infections in India, coronavirus may end up killing around 1000 people to 5 million people in India, according to the Epidemic Calculator. Why such gap between these two numbers?
According to the Epidemic Calculator, it all depends on flattening the curve, bringing down R0 the rate at which infections spread and at what stage the lockdown was enforced and ordered.
Calculations in a while but before that a few words about the tool. You can check the tool here (http://gabgoh.github.io/
At the time of writing, the coronavirus disease of 2019 remains a global health crisis of grave and uncertain magnitude. To the non-expert (such as myself), contextualizing the numbers, forecasts and epidemiological parameters described in the media and literature can be challenging. I created this calculator as an attempt to address this gap in understanding, notes Goh. This calculator implements a classical infectious disease model -- SEIR (Susceptible > Exposed > Infected > Removed) -- an idealized model of spread still used in frontlines of research.
By default the tool already has the relevant data, based on the research reports that have come out on Covid-19 from China and other countries so far, in the required fields. To see how Covid-19 will progress in India on the basis of this available data about coronavirus, and how it infects people, how many it kills, for how long it keeps people sick, all you need to do is change the population of the region to India's population.
Here are some projections this tool gives as far Covid-19 and its future progression in India is concerned:
Covid-19 in India with default data
With default data, which considers the infection rate R0 of 2.2 it means a coronavirus patient is likely to infect at least two more people the coronavirus Epidemic Calculator indicates that in a year the disease will infect a little over 1 million people in India. Of these, it will kill little over 20,000 people. While you look at these numbers, there are a few important bits to remember.
-- By default the Epidemic Calculator assumes that on the 100th day after the first reported case the government has forced the region or country into a full lockdown.
-- The lockdown or intervention in some other way, brings down the R0 to less than 1. This is very important.
-- If the lockdown or intervention has brought the R0 lower earlier, say just 50 days after the first case was reported, the total number of cases will be significantly less and number of fatalities exponentially lower.
-- It is also important to remember that by default the Epidemic Calculator considers fatality rate for Covid-19 as 2 per cent. In China, it has been slightly more. In South Korea and Germany it is significantly less. In Italy and Spain, it is alarmingly high, particularly in Italy where it is above 8 per cent. If you consider other figures for India depending on which one you do the covid-19 picture will change.
The covid-19 case fatality rate for India is something neither experts nor Epidemic Calculator can predict. The global fatality rate for swine flu is less than 0.1 per cent. But in India, in the previous swine flu epidemics, the death rate has been close to 6 per cent. For Covid-19 only time will tell what is the fatality rate in India.
Covid-19 in India a best case scenario
So, what is the best case scenario for covid-19 in India. If everything works well in India, if government mounts a good response, and death rate is not above 2 per cent, health care system holds up, quarantine is effective, then covid-19 may end up infecting a little above 15,000 people. This also takes into account the fact that in India the first case was detected on January 30 and the lockdown was ordered on March 24. Assuming that the coronavirus became infectious in India a few days before the first patient was detected, we can say that the country mounted effective response on around Day 65 of the disease. This can potentially limit number of Ccovid-19 cases in India to around 15,000 and deaths to around 300. Let's hope this is the case.
Covid-19 in India worst case scenario
Now, take the worst case scenario. This is the one where the data that we have so far from the government is not accurate, the quarantine is not effective, the health system is not ready and will not be able to hold up well, that the disease is spreading in India undetected for far longer even though the first case was only found on January 30. In this case, also assume that India's poor health system will not be able to cope up with Covid-19 when cases start to rise and let's move the case fatality rate to 4 per cent, which is still significantly lower than what Covid-19 is showing in Italy and Spain but is higher than the global rate. Let's also assume that the quarantine doesn't bring down the R0 much. May be a half per cent or so but not enough to flatten the curve in one year.
In this case, given that population of India is little over 1.3 billion, we may end up seeing around 180 million infected in the country this year and little over 5 million dead.
This is all projection
It's important to note that Covid-19 epidemic calculator is a tool that creates projections on the basis of data you feed into it. And as far as Covid-19 and novel coronavirus are concerned, right now not enough data, and not accurate enough data, is available. Everything is an approximation. What is the real fatality rate of Covid-19? Different countries are showing different numbers and they are all over the place, with extremely huge gaps between them. How quickly and how widely this is R0 that Covid-19 spreads? Again, we don't know the exact number. Some studies say it is 4. Some say it is 2. Some even say that it is 6.
Then there are other unknowns. Can Covid-19 re-infect people who were sick with it and then got better? We don't know for sure. Does temperature India is a very hot country and dry air summer is coming in parts of India affect this coronavirus? And if it does, how much? We don't know for sure. There are many more factors, like real world conditions. How long Covid-19 patients take to recover? For how long they are infectious? For how long they have to be in a hospital? For how long a country can care for its ICU patients? These are the figures that government experts and policy makers probably know. An online epidemic calculator doesn't. So if you are giving the online epidemic calculator created by Gabriel Goh a spin, keep this in mind.