4 ms·
I worked in data and business analysis for the Australian Red Cross Blood Service for several years, and I know the tables mentioned in the article well. I have
by siddboots 10y ago
I worked in data and business analysis for the Australian Red Cross Blood Service for several years, and I know the tables mentioned in the article well. I have spend many hours of my life worrying over the potential risk of a data leak, especially for medical questionnaire and medical test data. Not mentioned in the article are blood test results for HEPB, HEPC, syphilis, and so on, but I presume they were also in the leak. (EDIT: on learning more about the nature of the leak, this data probably was not in it)
An outsider might suspect that this occurred because ARCBS is the sort of organisation that simply doesn't have the speciality/budget/staff required to manage data security properly, but this isn't the case by any means. My retrospective impression is that they had capable teams covering data warehousing, data security policies, network security, etc. All of my experience in the field tells me that they were better protected than the vast majority of organisations of their size.
This event has solidified some of my more apocalyptic thoughts w.r.t. info-sec. Namely, that a guarantee of secure data is essentially out of reach for almost any organisation of any size. Partly this is due to increased connectivity of networks, but equally it is due to better tooling and analytics staff making thorough organisation of customer data the norm.
The cause of this problem is the hoarding of knowledge, as Zhuang Zhou realised two and a half thousand years ago:
> In taking precautions against thieves who cut open satchels, search bags, and break open boxes, people are sure to cord and fasten them well, and to employ strong bonds and clasps; and in this they are ordinarily said to show their wisdom. When a great thief comes, however, he shoulders the box, lifts up the satchel, carries off the bag, and runs away with them, afraid only that the cords, bonds, and clasps may not be secure; and in this case what was called the wisdom (of the owners) proves to be nothing but a collecting of the things for the great thief.
- x1798DE 10y agoThis is exactly why large stores of data should be considered a liability and NOT an asset. Keep as little of it as possible and for the least amount of time possible. I'm actually a bit confused about why they would need to store the medical test data and medical questionnaire information at all - can't you boil those down to "suitable" and "not suitable"?
- imglorp 10y agoOne would imagine a case where some blood product recipients got sick after a transfusion, and the data would be helpful to narrow down the cause relative to donor histories. I bet you could encrypt all the histories with a public key however, and only decrypt in such cases: private key held in a safe in the director's office, etc.
- jacquesm 10y agoYou are absolutely on the money here, however, as a rule it seems companies and their management feel that more data equals more power and potential income and so they are loathe to discard anything, even long after it has lost all value except to an attacker.
- nchammas 10y ago> This is exactly why large stores of data should be considered a liability and NOT an asset. Your comment reminded me of a post by Bruce Schneier with a similar theme, called "Data Is a Toxic Asset" [0]. [0] https://www.schneier.com/blog/archives/2016/03/data_is_a_toxic.html https://www.schneier.com/blog/archives/2016/03/data_is_a_tox...
- jacquesm 10y ago> All of my experience in the field tells me that they were better protected than the vast majority of organisations of their size. Which is a nice tell on what's going on elsewhere without having direct insight.