Generic domain question: Does anonymizing data decrease the data subjects’ ownership of the data?

[Oleks]

Does anonymizing data decrease the data subjects’ ownership of the data?

[bioran]

This is a leading question. Were it possible to "anonymise" data and leave it statistically useful then partial ownership/stakeholder questions would be easier to resolve in most cases. There is less often a risk of harm to the individual and so their consent is likely. That a patient becomes less of a stakeholder is never true.

In actual fact statistically useful (patient level) data is never anonymous. All are subject to Inference Attacks, with the ease depending on the effort an attacker is willing to spend on the available data fields. Consider knowing 5 attendance dates for a patient and then finding those data records with exactly that pattern of visit dates.

In summary because no generally useful data is anonymous (a falsely binary suggestion) re-identification is always a risk. How much? Well you have to measure that and then make your case to the populace.