- From: Alan Ruttenberg <alanruttenberg@gmail.com>
- Date: Sat, 3 Nov 2007 23:14:38 -0400
- To: "eric neumann" <ekneumann@gmail.com>
- Cc: "Kashyap, Vipul" <VKASHYAP1@partners.org>, "Forsberg, Kerstin L" <Kerstin.L.Forsberg@astrazeneca.com>, public-hcls-dse@w3.org, "Stanley Huff" <Stan.Huff@intermountainmail.org>, "Oniki, Tom (GE Healthcare, consultant)" <Tom.Oniki@ge.com>
On Nov 3, 2007, at 10:20 PM, eric neumann wrote: > A key item regarding and Adverse Event is that one does not usually > know at the time that a "adverse Event" is due to the drug being > administered. > > How the drug becomes "suspect" with the event is not always an easy > process, and sometimes it relies on accumulating many similar events > under similar circumstances from different subjects... Do you know, for the SDTM, whether adverse events recorded there have to be the ones that are determined to be related to the treatment, or are they events that are any events determined to be undesirable, reported to the FDA, and taken as input for later determination whether they are drug related? > Therefore an important logical constraint to weave in is: > At time t0 we see an event AE under the treatment condition Z, but at > time t1 we determine that the AE is due to (or not due to) treatment > Z. This fits well with a situation in which at t0 we talk domain (patient received dose of drug, stomach bleeds) and at t1 we add further statements that link the event with a cause, perhaps buttressed with domain knowledge such as mechanism of action. > I would also consider the use of "reasoning with uncertainty", but the > prospect of growing evidence an important feature to include for any > CO ontology... Have you see the latest pellet move in that direction? Might be worth playing with. If someone can come up with plausible examples, we could see about them coding something up or giving us access for some experiments. http://clarkparsia.com/weblog/2007/09/27/introducing-pronto/ http://clarkparsia.com/weblog/2007/10/02/using-pronto/ Best, Alan
Received on Sunday, 4 November 2007 03:14:50 UTC