COVID-19 Vaccination: What does it hold for India?

Introduction

The Covid-19 pandemic has affected countries all over the world and India has been no stranger to this. Starting from travel bans, to major shortages of protective equipment, hospital beds, oxygen cylinders, and even doctors, the people of the country have been through a lot of suffering. With the onset of the vaccination drive however, there seems to be a promise for a better future.

I was curious about how well the vaccination drive is working, so I decided to break the problem into a few smaller questions:-

Which 5 states should we choose to visualize vaccination curves?
Is there any difference in the administration of Vaccines with respect to gender, type of vaccine, or age group?
Given the above information, can we estimate the number of days it will take to vaccinate all individuals in a particular State?

To explore this, I gathered vaccination data from 5 states in India – Delhi, West Bengal, Maharashtra, Karnataka and Uttar Pradesh, and used R to visualize the data based on different categories. Using this data and the census data from each state, I tried to accurately estimate the time needed to vaccinate all individuals in that state. It is important to note however, that this model is based purely on vaccination data. If combined with additional data on infection and mortality rate of each state, it will be even more sensitive and hence yield a more accurate estimate.

Through this article, I will share my thought process, findings, and analysis.

Part One: Why choose these 5 States?

The five states I chose were Delhi, West Bengal, Maharashtra, Karnataka and Uttar Pradesh. The most important reason for choosing them is because all five of them are among the top 10 states in India to have the highest number of infections (1). Additionally, I thought it would be interesting to compare the rates of vaccination between states like Delhi and Uttar Pradesh, considering UP has a population that is 7.7 times that of Delhi.

Part Two: Visualizing Vaccination Data using R

The next step was to visualize the data gathered using R and observe patterns in the data. In addition to graphing the total number of individuals vaccinated with time, I thought it would be interesting to compare the 5 states based on 3 categories: Gender, Type of Vaccine, and Age. Let’s see how the states compare in each category.

A. Overall Vaccination Curve – Total Individuals Vaccinated with Time

Graphs of 'Total Individuals Vaccinated with Time' for each of the 5 states — **Graphs of ‘Total Individuals Vaccinated with Time’ for each of the 5 states**

In all 5 states, we observe a sigmoid or S-shaped curve. This pattern is characteristic of population curves, so it is no surprise that we see it for vaccination curves as well. The vaccination rates lag at first, but then increase exponentially after around day 50 (or March 7, 2021). Overall, it is good to see that all five vaccination curves show exponential growth!

B. Comparing the Curves with respect to Gender

Graphs of 'Individuals Vaccinated with Time with respect to Gender' for each of the 5 states — **Graphs of ‘Individuals Vaccinated with Time with respect to Gender’ for each of the 5 states**

The data available only includes three genders -Males, Females, and Transgender individuals. Although it is good to see that the curves are similar for males and females in all five states, it is also important to note that the disparity between the two genders increases with time. The only exception to this pattern is Karnataka. Along with being the only state where the vaccination curve for females is steeper than that for males, it is also the only state where the disparity between the two genders does not increase with time.

Although transgender individuals have been included as a part of the dataset, their vaccination curves appear to be straight lines near zero in all five states. The reason for this may be that the number of individuals who identify as transgender is very low compared to those who identify as males and females.

C. Comparing the Curves with respect to Type of Vaccine

Graphs of 'Individuals Vaccinated with Time with respect to Type of Vaccine' for each of the 5 states — **Graphs of ‘Individuals Vaccinated with Time with respect to Type of Vaccine’ for each of the 5 states**

In India, there are currently three vaccines that have been approved -CoviShield/ Astrazeneca-Oxford vaccine, Covaxin and Sputnik V.

In all five states, we see a huge disparity between the curves for CoviShield and Covaxin. It may be that people are intentionally choosing CoviShield over Covaxin; or perhaps CoviShield is more readily available and is able to meet the supply requirements for all states.

In all five states, we observe the curve for Sputnik V to be a straight line near zero. This suggests that the number of individuals who received the Sputnik V vaccine is far fewer in number than those who received Covaxin or CoviShield. This is expected as Sputnik V was just recently approved by the Indian government and was made available in the market in the week of 17th-23rd May (2). Hence, it is not a surprise to see that the curve has not grown exponentially yet.

D. Comparing the Curves with respect to Age

Graphs of 'Individuals Vaccinated with Time with respect to Age' for each of the 5 states — **Graphs of ‘Individuals Vaccinated with Time with respect to Age’ for each of the 5 states**

The CoWIN Vaccination Portal divides the population into three different age categories – 18–45, 45–60, and 60+.

Generally, we see that age groups 45–60 and 60+ have similar vaccination curves. However, this is not the case in Delhi since the vaccination curve for the 45–60 age category is far steeper than that of the 60+ category. Additionally, we also see that the curve for the 60+ age category starts flattening around day 80 (or April 6, 2021) in Delhi. The reason behind this is unclear.

In all 5 states, the vaccination curve for the 18–45 age group lags at first, but then sees massive exponential growth after around day 105 (or May 1, 2021). This makes sense since the CoWIN Vaccination Portal allowed 18+ individuals to register for their vaccinations only after May 1, 2021.

Part Three: Estimating the Number of Days it will Take to Vaccinate all Individuals in a Particular State

Now that I had visualized all this data and observed certain trends, one thought came to mind – is there a way we can predict the number of days that it will take to vaccinate all individuals of a particular state?

To answer this question, I adopted a simple and mathematical approach. Using R, I developed a linear regression model and acquired a trendline for the total number of individuals vaccinated with time for each of the five states. Then, using the equation and India’s Census data (3,4), I calculated the predicted number of days it would take to vaccinate all individuals in the state by working backward. My findings have been summarized in the table below.

Note: The above-mentioned population of Karnataka was last recorded in 2020, i.e., the numbers have not been updated for 2021 yet. For all other states, the population is based on current numbers.

The numbers above are quite interesting. My observations are summarized below:

Delhi has the smallest population out of all the states, and we predict that it will take 345 days to vaccinate all individuals.
In spite of Karnataka having a population that is 2.3 times that of Delhi, it predicts 342 days to vaccinate all citizens (3 days less than Delhi). This suggests that Karnataka is doing well with its vaccination drive.
Maharashtra’s population is 4.2 times that of Delhi and 1.8 times that of Karnataka, and yet it predicts only 355 days to vaccinate all individuals. This is just 10 days more than the predicted number of days for Delhi, which is quite praiseworthy given its population.
Lastly, we have West Bengal and Uttar Pradesh. West Bengal’s population is 3.3 times that of Delhi while Uttar Pradesh’s population is 7.7 times that of Delhi. In spite of such a huge difference, the predicted number of days to vaccinate all individuals in these states are 549 days for W. Bengal (just 1.6 times the predicted time for Delhi) and 585 days for UP (just 1.7 times the predicted time for Delhi). This is very commendable and suggests that W. Bengal and UP’s vaccination drives are working efficiently considering their large populations.

In summary, the estimated percentage of the population vaccinated after 30, 180, and 360 days are shown in the bar plots below.

**Bar Plots showing the Estimated Percentage of Population Vaccinated in each State after 30, 180, and 360 days respectively**

Assuming the current vaccination rate- Delhi, Maharashtra and Karnataka are expected to vaccinate all individuals in the next 1+ years whereas Uttar Pradesh and West Bengal will take a while to do the same.

Final Thoughts

With a population of 1.39 billion, India’s vaccination drive is one of the largest in the world. Experts say that 70–85% of the population will need to get vaccinated to achieve herd immunity, which translates to at least 973 million people (5). While this analysis only considers 5 states and data from a limited amount of time, it will be interesting to see how the vaccination curves change once more data is available.

This model takes into consideration data purely obtained from vaccination drives. If integrated with more information on the infection and mortality rate of each state, it will help us to relate vaccination drives with the infection rate. This will be a very important policy tool in the hands of the government and health officials to chart the territory to a COVID -free India.