Has somebody shared your 'anonymised' health data? Bad news

Researchers from Harvard University have published a paper claiming a 100 per cent success rate in de-anonymising patients from their supposedly anonymised healthcare data in South Korea. The study, which bears the ronseal title of "De-anonymizing South Korean Resident Registration Numbers Shared in Prescription Data", was …

[Your alien overlord - fear me]

Isn't the main worry that dead South Korean's are still getting drugs?

[Turtle]

Dead.

"Isn't the main worry that dead South Korean's are still getting drugs?"

Don't worry about it. The dead can take all the drugs they want and it won't hurt them at all.

[Zog_but_not_the_first]

Well, knock me down with a kipper!

Who'd 'a thought it?

[graeme leggett]

Source of IMS data

Glossed over is where, IMS get the data from. They don't just turn up at your Doctor's door with a wodge of cash and say "can I have a look at your patients records"

IMS sell the tools that pharmacies and surgeries want to use to organise the prescribing and dispensing of medicines. This gives them access to the prescribing and dispensing data straight off.

The pharmaceutical companies then pay IMS for beautifully processed and presented digests (eg product and presentation prescribed against what is actually dispensed by region over time) of the prescribing/dispensing data so they can see how/where their products are being used compared to their competitors.

That said, the ordinary person can go to the Health and Social Care Information Centre and see what GPs are prescribing -though not who to and what the pharmacist actually gives them. (http://www.hscic.gov.uk/gpprescribingdata) . You'll need to parse the data yourself but its there for the viewing.

[Anonymous Coward]

"Anonymised" data has always been questionable to the point of BS...

...for example, if someone's "anonymised" search history is reasonably complete, it's likely to include searches on themselves or people very close to them, the former of which is trivial to spot, the latter well within the bounds of unremarkable mining of publicly-accessible data like Facebook relationships.

While no single piece of evidence may be definitive proof, the patterns traceable in a record of any length can add up to much stronger evidence.

Join the dots and bingo! Bob's your uncle... and since we know Bob only has one nephew, you must be Allan Raymond Smith of 17 Finsbury Court, Scunthorpe, PP1 2TP.

[h4rm0ny]

Ask any IT professional in this field or related...

...Whether they think it's probable that this level of detailed data can be de-anonymised, and they will tell you "yes." Unless they work for one of the companies providing it.