Review Previously marketed, unapproved, withdrawn and discontinued

[grishagin]

https://drugs-dev.ncats.io/substances?facet=Development%20Status%2FDiscontinued&facet=Development%20Status%2FUS%20Previously%20Marketed&facet=Development%20Status%2FUS%20Unapproved%20Medicine&facet=Development%20Status%2FWithdrawn&facet=Substance%20Form%2FPrincipal%20Form&page=1

I'd like to propose reviewing RCAP status especially on these above categories of drugs and reviewing for accuracy for:

1. Initial marketing date/year
2. Current marketing status

I had earlier neglected to include 'previously marketed' into this list.

An initial logic review is probably warranted, to understand which types of events push drugs into these different categories.

Perhaps we can not have a rigorous separation between the different categories and should instead group them all together in 'previously marketed'.

Withdrawn is particularly problematic. We should update the logical threshold for withdrawn to be, all forms of the medicine removed from the market due to safety concerns (i.e. Bextra https://en.wikipedia.org/wiki/Valdecoxib) as compared to something like THIOBUTABARBITAL 2N0251U7JH. And of course, we should provide an annotation specific to the US, instead of the grouped way they are mashed together in combined_withdrawn_shortage_drugs.txt

Initial US marketing dates and references can be set by including that into FDA-NMEs-XXXXX.txt file in the data directory of stitcher.

Nutritional supplements (cysteine 48TCX9A1VT, sucrose C151H8M554, etc.) are particularly problematic. We should not maintain a database of all food and nutritional supplements, but only those that are used as drugs.