Closed wbryant closed 2 years ago
100% agreed. A lot of the push for this policy is related to our new data strategy which speaks to many of those challenges, and I sit within our Standards and Interoperability team. I'm very happy to reference both of those facts and our wider work right at the top.
One of the biggest challenges in sharing code for clinical research/analytics is consistent data models. This can be addressed by bespoke per-project data engineering or institutional adoption of a standard data model (e.g. OMOP/FHIR) or (more likely) a combination of the two.
In order to promote open tools, challenges around data standards and interoperability need to be addressed - this should at least be referenced here. It addresses the same "never the same job twice" advantage of open sourcing and indeed is a prerequisite for sharable analytics code.