.@drbeccawilson introduces another very current-affairs relevant topic at #CollabW20 - data privacy in health! how can we gather, use, analyse, and safely+genuinely anonymously share personal health data? 
Pseudonymisation makes data more useful, but comes with a hitch - linking it to other data sources can result in the data being de-anonymised #CollabW20 😱 @DrBeccaWilson
So, at #CollabW20 we've learned that anonymised and pseudonymised data is really hard to do usefully and safely. Data safe havens allow access to anon/pseudonymised data for research, but keep the data safe on secure remote machines. @DrBeccaWilson
Data safe havens still have restrictions and limitation - e.g. you can't just download your favourite script from GitHub to run your analysis! this also brings the risk of data silos, and you can't really stop screenshots. #CollabW20 @DrBeccaWilson
WOW - the most secure data sharing option: secure data pods - I never imagined this! A tiny box without networking, which you can enter wearing only your clothes (and may be video recorded while accessing data) #CollabW20
.@DrBeccaWilson Introduces DataShield, an R-based open source infrastructure for dostrobuted remote analysis! To the user, it looks like a standard R library use, but individual users can't see the individual level data - only seeing results at the end. #CollabW20
So - to wrap up - we need strong data governance for ethical and legal reasons. Doing the legal minimum often isn't enough for true safety, and most solutions have tricky downsides.
