Hi Melvin I suggest you have a look at the company Mmodal. They have gone a very long way toward clinically tuned speech recognition + NLP + populating and acting on spoken clinical data I am not directly connected with them but had the opportunity to meet their team and see the capabilities of their system.
Peter On 1/17/17, 2:29 PM, "Melvin Ma" <ma.qian...@gmail.com> wrote: >I am a programer / entrepreneur working on a voice data entry system for >EMR. To me, the future of EMR data entry will be voice based. I am still >researching / analyzing this idea and am happy to discuss with anyone >interested. Technically, the key technologies for my project will be >speech >recognition and natural language processing. I am currently intensively >working on the speech recognition part. > >My other interest is to create shortcuts for physicians to >expedite/automate their routine tasks. > >Before starting this adventure, I worked for Kaiser Permanente for close >to >10 years. In addition to kp.org/mydoctor web site and a video visit >application, I built a population management suit from the ground up. We >extracted data from various clinical systems, used domain specific >language >to describe care-path rules and calculate the reminders for all patients >in >the north California region. > >Qianfan "Melvin" Ma > >On Tue, Jan 17, 2017 at 1:20 PM, Thomas W Loehfelm ><twloehf...@ucdavis.edu> >wrote: > >> Now seems like as good a time as any for an introduction. It is great to >> hear about other organizations working to install cTAKES. >> >> I'm a radiologist at UC Davis in Sacramento, and came across cTAKES back >> in July. I have it up and running on a personal Windows laptop where >>I've >> used it for some small scale projects, and am working with a few other >> people at UC Davis to install it in a server environment where we can >>apply >> it to NLP tasks on larger clinical document libraries. My IRB for >>access to >> 10+ years worth of radiology, pathology, and operative notes should be >> approved any day now! >> >> I haven't quite figured out the final work flow. I'm debating between >> coming up with a default analysis pipeline, processing all of the >> documents, storing the CASes that result, and pulling subsets of them >>for >> information extraction on a project-by-project basis (storing >>information >> as needed in ad hoc MYSQL database schema) vs. custom processing >>pipelines >> for each specific project and using the built-in YTEX database as the >>main >> repository. >> >> Thanks to you all for your work so far in developing and sharing >>cTAKES. I >> am not very facile with Git, Maven, or formal software engineering >> concepts. I am willing to learn though, and willing to help out >>wherever I >> can contribute. >> >> Regards, >> Thomas Loehfelm, MD, PhD >> Assistant Professor, Radiology >> Abdominal Imaging >> University of California, Davis >> >>
