Metadata augmentation of @ausdataarchive datasets using @researchgraph augment API (proof of concept) - a big step forward for decision-support dashboard dev! Operationalising the #FiveSafes - one #SafeSetting at a time for CADRE platform. cadre5safes.org.au/2022/08/1…

The Future of Health Data Management: Enabling a Trusted Research Environment #TRE #SafeSetting #5Safes technologynetworks.com/infor…

You're not adhering to the #FivesSafes if you allow people's #PersonalData to be copied outside the #SafeSetting, and in making evidently #false claims to be following the #5Safes, @educationgovuk undermines public #trust in #TREs more generally. Which others should care about.

Hmm. Why is @HDR_UK hiring for 'Data Analytics & Research Environments': hdruk.ac.uk/jobs/senior-prog… when it hasn't delivered a #TRE* or shown why, despite not being one, it should tell #DataControllers how to manage #data access? _ *What ARE the #5Safes without a #SafeSetting? 🤔

#OpenSAFELY processes identifiable patient data in situ, using pre-approved code, providing only (non-disclosive) tabulated data as outputs. Which is good! A #FiveSafes #TRE has a bunch of 'layers' around the #SafeSetting, which is just ONE of the five: blog.ons.gov.uk/2017/01/27/t…

*sigh* It'd help if people would use the term '#FiveSafes #TrustedResearchEnvironment' for their #SafeSetting. Different bodies using different analogies for what they do with patients' #data really doesn't help public #understanding. Or public #trust. (Ditto "ownership" 🤦‍♂️)

Sorry - my bad! "#TRE" = "Trusted Research Environment" or, in other words, a #FiveSafes* #SafeSetting like, e.g. @ONS uses for the raw #census data. Like we first proposed... in 2014: medconfidential.org/wp-conte… _ *blog.ons.gov.uk/2017/01/27/t…

It is for others to determine the extent of #UKHRA's "leadership" on #TREs and the #FiveSafes. We would just note @medConfidential recommended @ONS's #SafeSetting to Parliament in 2014 during the #caredata debacle, 3 years before @HDR_UK was founded & 5 years before #UKHDRA...

It's well known many have concerns around commercial interests 👇 getting copies of NHS patients' data (one reason why we say patient-level data should never leave the #SafeSetting). See also: theysolditanyway.com/

The Government COULD have Directed @NHSDigital (which basically has to do what it's told) that #NHSpatients' #data should *only ever* be made accessible via a '#SafeSetting' or '#TrustedResearchEnvironment (like is used for the census data & National Statistics). It chose not to.

They'll be doing 'pseudo-at-source' but as we all know, that's just a sensible security measure; #pseudonymised data is still individually-identifiable #PersonalData, and @NHSDigital holds the keys. This time they do have a #SafeSetting - they just can't get customers to use it!

Update - Affinia is set up on lot at Delmar and Hamilton south side of street until 3pm today #SocialDistancing #SafeSetting #FreeCOVID19Testing @TheT5874Delmar @AffiniaHealth @lj_punch @storystitchers

Without knowing the precise details, in principle it seems similar to the way a #SafeSetting could work: you design and test your #query/#algorithm on #synthetic or #sample data; when approved it's then run on the actual data; and what you get back is the #nondisclosive results.

While @GenomicsEngland does indeed have far better policies/controls than @NHSDigital as regards individual-level data in general, @vivienneparry may be able to confirm (as she did to me a while back) that genomic data *does* in some cases leave the #SafeSetting. And further...

Uh-oh. #SingleSourceOfTruth never works out well, no matter who tries it. One way to stop "conflicting" data sets would simply be to stop selling / handing out copies of them all the time! We proposed using a #SafeSetting for patients' data back in 2014 - what news of that?

Great seeing @HAGIS_SCOT in @adrc_scotland safesetting Tks to Anthea & Amy support in this breakthru #ADRN2017 talk: teamup.com/event/show/id/rSk…