Breaking: UK government was routinely overstating the total number of people who’d been tested for #COVID19 by as many as 200,000 at the height of the coronavirus pandemic, according to new Sky News analysis.

This revelation stems from data released by @DHSCgovuk yesterday in the wake of our investigation into the handling of #COVID19 data. Full thing here: https://news.sky.com/story/coronavirus-the-inside-story-of-how-uks-chaotic-testing-regime-broke-all-the-rules-12022566

You may recall that up until late May @DHSCgovuk published daily figures for how many people had been tested alongside its deaths numbers. This is one of the key datapoints in confronting the disease and building a track & trace system. Here's the fig for May 21: over 2m

But on May 23 it abruptly stopped publishing the data. From then until this week we haven't been given any update on how many people in total have been tested in England (Scotland, NI & Wales continued publishing). It's been one of the longest-running mysteries of #COVID19

A mystery with real consequences, since data on people tested (as opposed to tests carried out) is v v important. Many people end up taking more than one test - so there's double counting there. Comparing cases to people is a crucial metric for judging the spread of the disease.

Then, yesterday, after @skynews's investigation, @DHSCgovuk quietly posted this data series, which finally ends the mystery. Bit rum. Still: put it alongside data for the rest of the UK and we can finally get a firmer grasp of how many have been tested https://www.gov.uk/government/publications/number-of-people-tested-for-coronavirus-england-30-january-to-27-may-2020

And lo and behold this new data series reveals that while @DHSCgovuk was reporting that around 2m people had been tested as of around 21 May, the real total was around 1.8m. It was overstating the reach of its testing effort quite considerably.

Here's a statement from @DHSCgovuk. Impression I get is they're sincerely doing their best to clean up a horrendously messy dataset thrown together in fraught circumstances early on in #COVID19. Tho is releasing this crucial data surreptitiously the right way to go about it...?

I had an initial stab at this data yday and compared apples to pears. Was saved some blushes by a contingent of dedicated Twitter data warriors inc @fascinatorfun @0liver_03 @jneill and many, many others who've also dug through the data and shared ideas. Thank you.