R-Value and it's Covid Travails

SARS-Cov-2 aka Covid-19, the virus behind the current pandemic continues to be a nightmare for everybody - especially for statisticians. 

It was bad enough to have the Chinese obfuscating the truth about the original 'Wuhan virus' last year, but now we have a range of mutants (and double-mutants) that seem to be spreading in ways that are difficult to predict. This is is making the work of those who prepare statistical models much more difficult, and this in turn is affecting the quality of decisions being taken by policymakers.

Take for instance, the reproductive-value or R-Value for Covid-19. R is the number of people that one infected person will pass on a virus. 

The lowest possible R-Value is R-naught or R0. Also referred to as the Basic Reproduction Number, it is the R-value of a population that is assumed to have zero immunity against a new infection. R0 is the division of the number of new infections by the number of existing infections, or the average number of new infections over an infectious period (R0= new infections/existing infections).

Measles has an R number of 15 in populations without immunity. It means that, on average, every 10 people with Measles will infect 15 other people. For Covid-19, the latest R-value estimate for India, is around 1.30. This number is showing wide variations across states with Uttar Pradesh recording the highest infection value at 1.62, followed by Andhra Pradesh at 1.61.

The problem with the R-Value is that it gives us a picture of the past - by the time you calculate it, the number is already outdated. This is because it is based on data recorded on hospitalisations and death figures to get a sense of how many people have the virus – but the problem with this is that, since the virus’s incubation period is so long, it only gives an accurate picture of a few weeks ago.

Add to this the complications coming from recent mutations of the virus in different parts of the world, each with its own incubation period - the "UK" B.1.1.7; the "South Africa" B.1.351; "Brazil" P1, "India" variants B.1.36 (Bangalore) / B.1.1.7 (Punjab) and B.1.617 which contains two specific mutations - E484Q and L452R which is capable of 'immune escape', and infecting those who have already been vaccinated. 

Add to all this other newer emerging variants of concern (VoCs) and we have all ingredients for a statisticians' perfect nightmare. 

---------------------------------------------

LINKS & REFERENCES

* https://www.bbc.com/news/health-52473523

* https://theprint.in/health/indias-covid-r-value-increases-to-1-30-this-week-the-highest-since-april-2020/633783/

* https://www.weforum.org/agenda/2020/05/covid-19-what-is-the-r-number/

* https://www.statista.com/statistics/1101713/india-covid-19-cases-by-type/

* R0 Explained - https://indianexpress.com/article/explained/coronavirus-r0-reproduction-number-explained-6354339/

Breathing Life into Statistics

There is lies, damn lies and statistics. And then there is Prof. Hans Rosling.

Rosling was Swedish specialist in Global Health, and a tireless evangelist for the cause of statistics - especially the art of presenting dull, boring numbers in stunning visualizations.

The first time I saw Rosling in action was in this TED video, made about 10 years ago:

Earlier this year, in February 2017, I was saddened to know that Rosling had died of pancreatic cancer. And today, for some reason, I found myself going through a BBC documentary on him, followed by an urge to revisit a few of the simplest statistical tools that I had learnt to use, but forgotten from disuse and disinterest, over a period of time.

While I could sail past the concept of Averages, and even wrap my head around the Normal (Poisson) Distribution, I had forgotten all about Standard Deviations (SD), Statistical Correlation and Dependence..

As explained in this video, the SD process consists of (1) collecting data (2) calculating the average/mean of the data set (3) subtracting this mean from each of the data units, and squaring the differences, (4) calculating, once again, the mean of the squared units, and, finally, (5) calculating the square-root of the 'mean of differences' to get the Standard Deviation, which tells you how diverse the numbers in your data-set are.

Why exactly do we go through this five-step process? What is the logic behind all this squaring and subtracting? I still don't know...but it does remind me that I had asked the very same question in a classroom many years ago, and failed to get a satisfactory answer. Then, as now, the whole thing seemed like a form of gymnastics with numbers.

And yet, Hans Rosling seems like the sort of guy who would have given me a clear, logical answer, and triggered in me a passion for data. How wish I had come across such a teacher, a few decades ago!

-----------------------------------------
LINKS:

TED Video - "The Best Statistics You've Never Seen!" - https://www.ted.com/talks/hans_rosling_shows_the_best_stats_you_ve_ever_seen?utm_source=tedcomshare&utm_medium=referral&utm_campaign=tedspread

Don't Panic: The Facts About Population - https://www.youtube.com/watch?v=FACK2knC08E

Gapminder Foundation - http://www.gapminder.org/

Obituary (Feb., 2017) - Hans Rosling: Data visionary and educator dies aged 68 - http://www.bbc.com/news/world-europe-38900572