All the data you can eat
Don't MissMy Corona

Here’s what happened to South Africa’s missing Covid-19 data

Adam Oxford
Adam Oxford
2020-03-252020-04-01

There’s a potential problem emerging in South Africa’s ability to the spread of Covid-19 around the country: poor quality data.

Like many of those in South Africa who’ve been trying to stay on top of government numbers about the spread of the coronavirus, we’ve been a little confused over the last few days by the sudden inconsistency in official reporting on the subject. Last week the National Institute for Communicable Diseasess settled down into a sub-optimal, but workable, daily PDF update in a format we were able to scrape, supplemented by some numbers that appeared in an updated graphic.

Not ideal, but rich in information about location, age, travel history and gender of every case. Then, on Monday, the wheels came off and updates became sporadic, on different channels (Twitter – personal and official – or WhatsApp) and lacking in detail beyond headline case numbers. And there are already inconsistencies with the numbers reported on NICD’s official dashboards.

Why, we wondered, can’t we just get access to a simple database or CSV file?

Last night, health minister Dr Zweli Mkhize took to Twitter to explain what’s going on, in a Tweeted JPG (…) statement. According to the minister, once a case has been confirmed, patients are supposed to fill in a form with their personal details and the information of those that they may have come into close contact with – and despite carrying a penalty of up to 10 years jail time if not returned, by the sound of it these aren’t coming back properly completed, and NICD is struggling to fill in the gaps.

The process described by Mkhize is that anyone on the contact list of a confirmed case must also be tested.

The kicker, and far more serious than the gaps in the data that are emerging, is that Mkhize signed off with a threat. Not just jail time, but if your forms don’t come back correctly completed, you will be publicly named and shamed so anyone who knows you can get tested.

For what it’s worth, I personally think this is a terrible, terrible idea. It might meet with high public approval, but there’s just too much history of mistaken identity and mob justice when it goes wrong – not just in South Africa but around the world.

Adam Oxford
  • Adam Oxford is a freelance journalist, media consultant and civic tech enthusiast. He also works closely with startups developing solutions to access to justice problems.

2 Comments

  • Alan Levin

    I am so concerned about lack of and inaccurate data provided by the government. Besides that I do not agree with you stating “your opinion of terrible ideas”, I also do not understand how the “forms” with contact lists has anything to do with “the missing data” …. the links are broken in this article… this is an awesome blog, I suggest a rewrite or something…. I personally believe that test data about age, location and gender should be live and public! that’s not hard and nothing to do with forms…

    • Adam Oxford

      Hi Alan, Thanks for the feedback – sorry about the broken links. I think NICD has taken down its map by the looks of it, so I’ll remove that now.

      This is the link to the second part of the statement which makes it clear that the relevant forms weren’t being completed properly:https://twitter.com/DrZweliMkhize/status/1242535906151673857/photo/1 (and given the number of revisions we’ve seen in the last few days and Tweets from NICD about cleaning data I don’t think the problem has gone away.

      Perhaps I wasn’t clear – anyone listed as a recent contact of a confirmed case should be tested according to this statement, but the contact lists themselves weren’t being properly supplied, and nor were the diagnosis forms by the looks of it. It’s a bit of reading between the lines, but like I say, it doesn’t look like the problem has gone away.

      The terrible idea (in my opinion, which I’m definitely going to share :)) is naming and shaming, not the data collection. That always, always ends badly.

      I completely agree with you about the need for regular, accurate and detailed updates. Yesterday we got different numbers from the DoH and NICD published within five minutes of each other. Yes, collating and cleaning this is hard, but with so many tests being done there has to be a consistent daily reporting time, or any professional trying to understand the data is going to working half blind.

Leave a Comment

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