I am a Standards Architect specializing in Healthcare Interoperability Standards Architecture in Interoperability, Security, and Privacy.

53070
Joined June 2008
Considerations for Implementing the SLS RI Implementing an SLS to do data tagging, so that fine grained Consent can be supported, can be done in many different ways. A systems design approach is needed. healthcaresecprivacy.blogspo…
2
37
Sensitive data as Venn diagram In Healthcare, Normal data is all data that is linked to an identified Patient and not specifically sensitive. Any data that is sensitive would be Restricted. healthcaresecprivacy.blogspo… – at Oostburg, WI
1
40
John Moehrke retweeted
We're going around the Moon. Come watch with us. Artemis II's four-astronaut crew is lifting off from @NASAKennedy on an approximately 10-day mission that will bring us closer to living on the Moon and Mars. The launch window opens at 6:24pm ET (2224 UTC). xitter.cf/i/broadcasts/1jxXgeyMk…
5,525
28,209
5,144
106,522
17,162,426
Word of the day: Connectathoning. When test partners work well together to get through an issue. – at Oostburg, WI
1
21
Two months ago, while working on my Empire game with AI, I had that quicksilver experience. When you push on a blob of mercury it slips out in some random direction. Every time I added a new feature some older feature would shift behavior. This was true even after I added unit tests and acceptance tests. The AI always took the path of least resistance on the current feature, and was willing to sacrifice older features. It would change tests, including acceptance tests in order to get the latest feature done. Telling the AI not to do that was ineffective. AIs are stochastic, and so are any rules you feed them. Rules _bias_ their behavior, but do not absolutely constrain it. When I call them out on breaking rules, they apologize and swear they won't do it again; but they can't really make that promise. They are, in the end, liars and cheats. The solution is to massively overconstrain them. Force them to write so many tests that changing a tested feature breaks many tests. They feel that force and retract the change. It's like peer pressure with a lot of peers. At the same time I reduce the chances for collateral damage by continuously forcing the AI to partition everything into small decoupled units. That way it is not easy to break one feature while implementing another. It also keeps the AI from getting confused by it's own messes. The final goal is semantic stability in the face of continuous development. The things that worked before keep working as they were; while newer things get added. This is a continuous effort. Acceptance tests, unit tests, TDD. Crap analysis, and mutation tests are run after a reasonable batch of changes and are tasked with reducing crap below 8, covering any untested behavior, and killing all surviving mutants. The size of the batch of changes is a judgement call. Too big and the analysis and repairs take a long time. Too small and the verification effort overwhelms the development effort. Side note. The mutation tests consume massive amounts of computer power. My cores are running full bore all the time; and that's even with differential mutation. There is something poetically just about this. The AIs require a massive amount of computer power to create. What they create for us takes a massive amount of computer power to keep stable.
25
27
6
277
20,933
Consent about AI There are various use-cases where a Patient might consent or dissent to various uses of AI. A patient might consent to use of AI for clinical decision support A patient might deny use of their data for training of AI. healthcaresecprivacy.blogspo… – at Oostburg, WI
37
Security Labeling Service - Reference Implementation - an AI vibe experience I have vibe coded a SLS and made it available open-source. Details and links on my blog article healthcaresecprivacy.blogspo… – at Oostburg, WI
1
1
69
Pseudonymization, the terms 'anonymization' and 'pseudonymization' described distinct technical methods for de-identifying data. Driven by regulations like GDPR and court decisions, the focus has shifted from pseudonymization as the method to pseudonymized is the dataset itself. – at Oostburg, WI
1
1
1
58
Key is who possesses the re-identification method. This subtle change has profound implications. healthcaresecprivacy.blogspo… – at Oostburg, WI
1
35
The third kind of contract I’m well-suited for involves working directly with product developers—whether client or server-side—to ensure their solutions optimally leverage existing standards. This role is critical for real-world adoption. healthcaresecprivacy.blogspo… – at Oostburg, WI
45
Healthcare Exchange Standards: Standards Development Contracts - healthcaresecprivacy.blogspo… – at Wisconsin, USA
1
45
Healthcare Exchange Standards: Small contracts healthcaresecprivacy.blogspo… – at Lancaster, WI
1
29
Starting small. Mostly boomers. #GoodTrouble – at Sheboygan, WI
1
2
66
We are able to interconnect so seamlessly on the Internet in large part thanks to Internet standards. Internet standards allow us to communicate easily, across a wide range of devices, providers, and systems. Learn more about standard setters’ unique role in running the Internet: internetsociety.org/blog/202…
7
9
909
This Monday morning started differently—I’m awake, ready to work, but with nowhere to report. healthcaresecprivacy.blogspo… – at Oostburg, WI
27
How to record that the Patient authored #FHIR Resources or elements healthcaresecprivacy.blogspo… #Provenance
1
2
99
IHE IT Infrastructure Technical Committee has published minor (Version 1.4.0) update to FormatCode vocabulary IG - profiles.ihe.net/fhir/ihe.fo… – at Oostburg, WI
20
IHE IT Infrastructure Technical Committee has published Mobile Care Services Discovery (mCSD) version 4.0.0 - mailchi.mp/ihe/ihe-iti-tf-su… – at Oostburg, WI
19
IHE QRPH domain releases for public comment Computable Care Guidelines (CCG) mailchi.mp/ihe/ihe-qrph-tf-s… – at Oostburg, WI
1
55