This is a nice little paper on problem lists, doing bit of simple but interesting analysis of what goes into.
The humble little problem list is usually about 3-6 brief lines in a typical hand-written one to three page hospital admission document. It's pretty much what the doctor or nurse writes down after they've asked you "do you have any medical problems?" I'm aware of at least five disparate system where problem lists are generated and stored on my hospital campus, and I haven't even gone looking.
If we could stick them together and "wiki" edit them that would be pretty grand. Sprinkle in a little "spell-checking", maybe some clever filtering and grouping.... maybe some automated interpretation of laboratory investigations... a little bit of decision support...
So what seems superficially like such a simple notion of combining a few lumps of text in a wiki, is actually an open ended challenge of artificial intelligence and lexical processing. That doesn't mean creating a nice "problem list aggregator" is not a low hanging fruit. An initial "dumb" system would be good step forward, but once that fruit is picked there are plenty more on the tree.
(Addendum: Ooooo.... someone else is singing my song!)
The humble little problem list is usually about 3-6 brief lines in a typical hand-written one to three page hospital admission document. It's pretty much what the doctor or nurse writes down after they've asked you "do you have any medical problems?" I'm aware of at least five disparate system where problem lists are generated and stored on my hospital campus, and I haven't even gone looking.
If we could stick them together and "wiki" edit them that would be pretty grand. Sprinkle in a little "spell-checking", maybe some clever filtering and grouping.... maybe some automated interpretation of laboratory investigations... a little bit of decision support...
So what seems superficially like such a simple notion of combining a few lumps of text in a wiki, is actually an open ended challenge of artificial intelligence and lexical processing. That doesn't mean creating a nice "problem list aggregator" is not a low hanging fruit. An initial "dumb" system would be good step forward, but once that fruit is picked there are plenty more on the tree.
(Addendum: Ooooo.... someone else is singing my song!)
No comments:
Post a Comment