Jon Massey

RE: https://mamot.fr/@pluralistic/116056785239179997

Love to see @pluralistic 's enthusiasm for OpenSAFELY! Few (friendly) corrections: We're the Bennett Institute for Applied Data Science at the Nuffield Dept for Primary Health Care Sciences - Ben is affiliated with Jesus but the rest of us are not. The patient data aren't held by the NHS trusts, but by the two major GP software providers (who have an effective duopoly over the English market). "modified MySQL syntax" slightly hurts my soul - ehrQL is intentionally not SQL-like ...

Feb 13 at 11:04amWeb
Show threadJon MasseyFeb 13
@pluralistic there's a great blog post on its design here https://www.bennett.ox.ac.uk/blog/2023/09/why-ehrql/ (there's another fantastic one on how we test it!). Publishing of results is not automatic, and certainly not after a set period. Researcher's analyses can only ever produce aggregate results and are subject to strict statistical disclosure controls. Two independent reviews of every single research output verify these before data are released! This rigour is a big part of our privacy stance.
Why ehrQL? | Bennett Institute for Applied Data Science

Why did we create a new query language?

Show threadJon MasseyFeb 13
@pluralistic all of the code that runs the platform is here https://github.com/opensafely-core (including ehrQL) and fully open source