CEO at Clinithink Ltd, physician, computer scientist, innovator, entrepreneur passionate about technology-driven change with people at the centre of the process
@tomaspueyo @rafa1982rafa @davemcclure Well please keep doing it. If you automate it to lift global data from e.g. JHU github to show global version of chart 2 you wld have done whole planet a big favour. Chart 2 is the single most useful visualisation I've seen that tells us how much control authorities have, or not.
Dr. Kingsmore and Michelle Clark @RadyGenomics are providing a talk entitled "Utilizing Machine-learning and Clinical Natural Language Processing to Diagnose Genetic Diseases in Record Time" on May 3 at 11:30 am at the Frontiers Conference. #CNLP https://t.co/oaZyncb4xo
@EricTopol@ScienceTM@RadyGenomics@radychildrens We @clinithink are thrilled to have been a collaborator, providing the #AI#NLP solution that enabled the automated phenotyping, complimenting WGS. The extraordinary innovation in sequencing needs to be matched with advances in phenotyping for patients to get maximum benefit.
Big problem for EHR based research: upcoding. Huge issue-just heard about large academic study that was messed up bc many patients misclassified as condition A (comparatively high reimbursement mental health condition) vs condition B (lower reimbursement)
Interesting data about human collaboration with computer..
Overall cancer diagnosis error rate:
Human: 3.5%
AI: 7.5%
Human + AI together: 0.5%
via @ianbremmer
Totally agree @john_chilmark@peterneupert. People, pathology, care processes are all messy. This stuff doesn't fit in drop_down boxes. Patients tell their doctors stories, drs write them down. Consumer tech (Alexa, Twitter, etc) is narrative, healthcare tech needs to embrace
@john_chilmark Messaging standards can only be prescriptive for highly structured information. H/C data is indeed "messy" - disproportionate amount of value/signal is in narrative text, even short snippets in message segments. NLP is essential to exposing the value in the data @Clinithink
@magillett@KentBeck @RichRogersIoT I find that the opportunity to learn is a privilege that, er, keeps presenting itself. I'm not sure the opposite trade exists: "If I promise not to learn/discover stuff, can I have a plan that runs like clockwork?". I wish.
Having worked in health informatics for many years, improving healthcare with the use of technology is something I am passionate about.
I am so proud of the part our solution has played in the speedy diagnosis of rare genetic disorders.#DataSavesLives.
https://t.co/FHzoC2sbjA