I just read a reliable article on the basis of scientific research methodology. Well explained, good examples, I enjoyed it. I'm med student, passioned about statistics, I translate numbers into words, results into sentences, and an idea into a probable reality. I would be delighted to follow you! Hope you can do the same!
I'm trying to wrap my head around this. It's a thorough treatment and makes some good points. But are you pointing out that they proposed an aggregation for model training without also specifying any way to protect the patients whose data ends up baked in (I think it's implied within the expertise of the supplying agencies)? Or that this would introduce a competing ruleset against which consuming entities would have to measure their own risk and compliance?
Hey James, first thanks for reading and taking the time to comment. Now to the question: my article uses the Nordic plan as a framework to discuss the differences in de-identified data definitions across oceans. Also it’s a great way to discuss the nuances of medical databases in general. I’m not suggesting the Nordic group can’t do it. But I am suggesting that it has GDPR consequences for users of the data since the de-identified definition changes depending on which side of the Atlantic you are on. Does that help?
I just read a reliable article on the basis of scientific research methodology. Well explained, good examples, I enjoyed it. I'm med student, passioned about statistics, I translate numbers into words, results into sentences, and an idea into a probable reality. I would be delighted to follow you! Hope you can do the same!
I'm trying to wrap my head around this. It's a thorough treatment and makes some good points. But are you pointing out that they proposed an aggregation for model training without also specifying any way to protect the patients whose data ends up baked in (I think it's implied within the expertise of the supplying agencies)? Or that this would introduce a competing ruleset against which consuming entities would have to measure their own risk and compliance?
Hey James, first thanks for reading and taking the time to comment. Now to the question: my article uses the Nordic plan as a framework to discuss the differences in de-identified data definitions across oceans. Also it’s a great way to discuss the nuances of medical databases in general. I’m not suggesting the Nordic group can’t do it. But I am suggesting that it has GDPR consequences for users of the data since the de-identified definition changes depending on which side of the Atlantic you are on. Does that help?