Zoom Logo

DHSI Grand Rounds - Shared screen with speaker view
Harold Lehmann
23:55
Even if pre-market ADEs are "easier," clinical trial reports often make it difficult to establish ADE *rates* (or other evidential measures)
Leigh Smith
26:12
I can still hear him
Chenyu Li
26:15
I can
Harold Lehmann
26:31
Ah, sorry
Harold Lehmann
29:44
Since may ADEs are rare, doubling the numerator (duplicate cases) doubles that rare rate
Harold Lehmann
45:09
I hope y'all see that Dr. Botsis' Zoom background is based on thee visualization slide.
Harold Lehmann
46:55
Drugs@FDA: https://www.accessdata.fda.gov/scripts/cder/daf/
Harold Lehmann
01:01:33
Please type any questions you have. You can also ask by audio/video.
REM
01:03:19
Taxiarchis, (c. all):Thanks for the insights about your successes with the project and with NLP. I am a fan of NLP, but it is clear that drug adverse-event reports are much more challenging than pathology reports. :-))Bob Miller--
Leigh Smith
01:04:56
More of a comment -- This is very inspiring. My mind is going wild with other applications of NLP of labels
gkim9
01:07:54
On the interface, how would an investigator extend a search along a different dimension?
gkim9
01:11:21
I can't get my microphone to work without screeching. What I was trying to ask was: in such an interface the invitation to me is to fish along different semantic dimensions that may not be readily available on what is displayed.
gkim9
01:17:40
Does this go into PREMEDLINE? I am reminded of the Hexamethonium problem that occurred (actually at JH) where the data was in the pre 1966 listings?
gkim9
01:17:57
Thanks