FDA: Approval Documents

[chackoge]

For the community of researchers interested in scientific impact of basic biology research, a walk 'backwards' from a drug to the basic research that enabled its development is of great interest. Literature references cited in FDA approval documents are not easily accessible. In this example, they are scattered throughout the document in footnotes instead of being in an Appendix. Further, this document is in PDF. Providing a list of cited references in CSE format as plain text would be of help. Including a tagged unique identifier such as a PMID, WosID, ScopusID, doi: would be much much better., e.g., 1234567.

[dportnoy]

@chackoge thanks for making the request! Sounds like there are multiple components to it:

1. Aggregate references in a structured way. Perhaps a single table, rather than being scattered throughout long form content.
2. Include a unique identifier (such as a PMID, WosID, ScopusID, doi) that could be used for linking and analytics.
3. Save the document in a more open, machine-readable format, such as plain text or HTML.

Did I get that right?

Have you seen this type of referencing done well elsewhere? I think having a concrete example of a good implementation should help the discussion.

[chackoge]

@dportnoy

a) aggregating references in a structured way is useful- some documents have all the references in Appendix 9B as I recall but the FDA document is still in pdf. Providing a machine readable list of unique identifiers would be very useful, i.e., every citation would have to have at least one of PubMed ID, ScopusID, WosId, or doi. Ideally PubMed since access does not require a subscription but not all citations are to be found in PubMed.

b) re: Have I seen this type of referencing done well elsewhere? Unfortunately not. I am attaching a file that I manually assembled that is highly derivative of Alex Pico's work.

seedset_csv.txt