@IAmSamFin If I understood your query, then a shouldn’t these fit into your scope? 1. https://t.co/HV1QXbJRlw 2. https://t.co/IiXCB9xwFN Still missing another one that I read a few months back. Hope the above fits your query.
@arumshisky @SeeTedTalk @evanmiltenburg @annargrs @stevebarolak @emilymbender @dirk_hovy @SteveMeystre @NLM_LHC Some approaches use a cascading regex/"black list"/ "white list" approach instead, e.g. Philter @atulbutte https://t.co/lMwtIHebsq
RT @atulbutte: We manually deidentified 4500 notes of 100 different kinds (out of our 70 million), used 2500 to build and 2000 to test. Two…
Great strides for NLP and use for research with clinical free text notes
Incredible new work!! This is huge - open source approach by @UCSFMedicine will accelerate responsible open medical data sharing for research/education. @StanfordAIMI hopes solutions like Philter can de-identify medical imaging text reports for our ope
Largest set of deidentified clinical notes disclosed to date, perfect for #nlp secondary analysis 👇
Awesome. Looking forward to giving it a try!
Philter runs in the safest of environments, can run on an air-gapped machine, no internet access needed. Tested on our corpus and the 2014 i2b2 test corpus, compared with the great tools Physionet and Scrubber. https://t.co/XVEsWeZ51q https://t.co/Fp6fYCpt
We manually deidentified 4500 notes of 100 different kinds (out of our 70 million), used 2500 to build and 2000 to test. Two annotators per note. This might be the largest set of deidentified notes disclosed to date (@UCSF IRB approved). https://t.co/XVEsW
THRILLED to announce Philter, our new tool to deidentify clinical text. Now more than ever, safe & responsible sharing of #EHR data is needed to fight pandemics & disease, but text is hardest to share. @BeauNorgeot built a new tool w/ a great team!
The manuscript for our new tool, Protected Health Information (PHI) filter ("Philter" for short) is now live: https://t.co/EdangZn4Gb Philter accurately and automatically removes PHI from free text clinical notes, which can facilitate their availability fo
Protected Health Information filter (Philter): accurately and securely de-identifying free-text clinical notes https://t.co/i91r4e8aLN