Neil Walker has made an important contribution here. While a good data scientist may think he has made some points that are blindingly obvious, the sad truth is that there is a very large, and potentially very influential, group of people who either don't know it or wilfully ignore it.

The official citation: Walker, N., (2017). All or Nothing: The False Promise of Anonymity. Data Science Journal. 16, p.24. DOI: http://doi.org/10.5334/dsj-2017-024 

The seminal point in this paper:

"some researchers and policy makers have conflated the notions of de-identification and anonymity. The former is a process that seeks to mitigate disclosure risk though careful application of rules and statistical analysis, while the latter is an absolute state". As Neil goes on to state in the paper, "The consequence of confusing the process and the state is profound".   

He very pointedly includes the International Committee of Medical Journal Editors (ICMJE) among those who have fallen for the fallacy.

For some practicl enlightenment on how to navigate this challenge, see:

"The Anonymisation Decision-making Framework" from the UK Anonymisation Network (with endorsement from the UK Information Commissioner) at: http://ukanon.net/ukan-resources/ukan-decision-making-framework/

AND the simplification and adaptation of the UKAN Framework: "De-identification Decision-Making Framework" written by the Data61 group in CSIRO & the Office of the Australian Information Commissioner at: https://www.oaic.gov.au/agencies-and-organisations/guides/de-identification-decision-making-framework.

This area needs a lot more discussion and even more education and skills development in the health research and health policy community.

Malcolm Crompton Lead Privacy Advisor, Information Integrity Solutions Pty Ltd Privacy Commissioner of Australia 1999-2004