OPINION: Credible Covid-19 recovery statistics are illusive
By Pali Lehohla
JOHANNESBURG - In 2014 I was part of the 25 member team to advise the UN Secretary-General on Data Revolution ahead of the UN General Assembly adopting the Agenda on Sustainable Development and the attendant indicators.
By November 2014 we tabled our report titled, The World that Counts to Ban Ki-moon. We raised major issues in the report, amongst these was the ease with which today the production of data has become leading to the era of Big Data.
Society is exposed to censors that gather transactional information data and transmits it instantaneously. Mobile devices have liberated society from fixed line phones. They have created on-line hand held capability to communicate, transact and learn.
The Ebola epidemic that visited West Africa could be contained with access to Orange Mobile data holdings and capability.
In Europe household surveys such as the labour force surveys are now individualised as capacity to target the individual is enabled by the existence of these mobile devices.
Women in Lesotho, who up to 20 years ago relied on a husband working on the mines to authorise farming activities by means of a letter, today can respond to weather instantly.
They can consult by phone and make decisions of whether to plough, plant or weed, without having to wait weeks on end to receive the letter.
With voice on devices, limitations imposed by illiteracy are overcome and there is no need to get someone to come and read the letter.
So ubiquitous is information technology, one wonders what the situation would have been if we did not have these devices under conditions of the coronavirus pandemic.
The devices have led to enhanced citizen participation in trying to understand the evolution and progression of the pandemic.
The combination of the devices and a dreadful high velocity disease have led to the visibility of data and statistics.
Both the pandemic and Big Data, which constitutes the observations of the pandemic, have come together under the common name of size and velocity.
Coronavirus within a wink of an eye had engulfed the world and thus fit the notion of Big Data, which in the main is unstructured observations that possess unprecedented size and speed as well as variability.
As a statistician and demographer, I always would like to put structure to observations and Covid-19 data has remained rather ever illusive.
By design and observation the data does not lend itself to population level estimates because the observations are not random and consequently they have an unquantifiable factor by which they could be raised to the universe to arrive at credible estimates of prevalence, evolution and progression.
The most intriguing number is one on recoveries.
But once a specified tested population the outcomes are known, either they are positive or negative.
If they are positive they have to self-quarantine and or subsequently go for medical advice. If they are negative, they are negative and they have to continue acting with deliberate care against infections.
Those who are positive become a new denominator, which should have a numerator of self-quarantine and hospitalisation.
These two components of the numerator have recovered, in progress or dead.
My family members tested Covid-19 positive and went under self-quarantine. After 14 days I had wanted the tribe to go and test, but I got mixed messages on this. They have not gone to test to confirm recovery, but my observation is that they have and the medical literature suggests that they should have.
But how then does the statistics of recoveries get computed if there is no reconciliation of the three components of the numerator?
The only credible number in the numerator is the number of people that died given that they were Covid-19 positive, because the dead ID number can be reconciled back to when it was alive and tested.
The number of people continuing to self-quarantine outside a health institution is impossible to know and separate from those who recovered because there is no process like the occurrence of death that forces such reconciliation.
The invisible enemy pandemic is true to form in its invisibility, and like Big Data, its data is unstructured, big with high velocity and generating rather illusive statistics.
Dr Pali Lehohla is the former Statistician-General of South Africa and the former head of Statistics South Africa. Meet him at www.pie.org.za and @palilj01