With Google and Amazon pushing into the healthcare space and collecting lots of available patient health records, I was wondering if it's possible to make that data open source so scientists and technical people all over the world can work on solutions, instead of only for-profit companies.
I've been wondering if it's necessary to build a company to do it - and your answers would greatly inform my decision and path to start one.
The following are the questions I've been pondering about - and would love your feedback:
- Nomenclature and data formats: how hard is it to develop a common language for many different, complex data formats with different naming conventions across countries? What exists already, what has been tried before?
- Are academic institutions (think Johns Hopkins University) allowed to share datasets publicly? I know the answer is 'Yes' since lots are already doing it, but what are specific (technical and legal) barriers to consider when doing this at scale and developing one platform?
- What are generally big challenges and problems for researchers in the medical sciences, or anyone else that professionally works with patient health data?
- I know data privacy is a big challenge - no secret. But does anyone have experience with specific technical and legal issues when it come to anonymisation, publishing data from secondary sources and their legal compliance? Please share!
Anonymizing the data does not work, as you can infer and rematch against other databases. Even encrypted data is susceptible to reverse engineering