Two questions! 1. Does anyone have extended experience of speech recognition for dictation of radiology reports 2. What is the current state of play with integration of speech generated text reports integrating with RIS (and then PACS)
posted on Friday, October 25, 2002 - 12:46 pm
Stephen You may find the SCAR site useful for this, you can link to this from our site. There was a presentation from Langer at the last SCAR meeting reported in J Digit Imaging which may be v useful (see refs to publications which I posted recently) Keith
At the Norfolk & Norwich we have been experimenting with voice recognition for nearly a year now. We are using Dragon Naturally Speaking which has been embedded into our RIS by the Dragon (GHG) rep. In practice we speak the Radiology number into RIS (no direct interface with PACS worklists yet) and separately launch the PACS image. At the end of dictation the report can be authorised and printed with one command. The printed copy is then posted off to the referring clinician and the electronic version is immediately available on the Reports and Requests systems and on PACS. So far two radiologists have gained substantial experience with it and like it. Three more are in the throes of learning and the majority remain sceptical but interested. Complete integration with PACS worklists would make the system much less "clunky" and aid acceptance by speeding up the whole reporting process.
Gerry, we maintain a user database which can be accessed from the link at the top of each webpage. You will need to click list options and choose which fields you want to display - including PACS, RIS, voice dictation vendors and Hospital/Institutions in your case. You can then contact the relevant member for more information.
I have been using Phillips SpeechMagic for about a year. It is fully integrated with the Amersham (Torex) RIS and works extremely well. I am hoping to do some live demos in the near future (eg at Calderdale). Contact me if you want more information.
Hi - Robert George, Adelaide Australia. Re - Speech recog with PACS .
There is an excellent site in Sydney Australia at the Sydney Adventist Hospital which has AGFA PACS and Speech recog using Talk Technology, recently acquired by AGFA.
I think they use the Phillips speech engine but you could contact the Radiology Manager, Geoff Andrews - see email below The Hosp web site is www.sah.org.au - go to Diagnostic Services This is a superb Hosp with a full cardiac facility etc
Geoffrey Andrews Radiology Manager 185 Fox Valley Road Wahroonga, NSW, 2076
The main site I know is Christie Hospital Manchester who have speech recognition with an Amersham (Torex) RIS for many years. Also possibly Central Middlesex. In Europe G2speech is interfaced to PACS in the Benelux (see G2 web site). Simon
We use Philips SpeechMagic with the Torex RIS. It is very successful, full integrated, and we love it! I recently demonstrated the remote reporting set-up we use at home by linking a laptop via ISDN from three different sites (London, Halifax and Warwick) to our Grimsby site. I downloaded images, did (dummy) reports using speech recognition and then corrected and authorised the reports. We use Radworks and link using ISDN 2 through Cisco routers, which is very secure.
Christie Hospital Manchester had speech recognition with Amersham (Torex) which worked well until 3 years ago when we upgraded to the windows version. We are still waiting to get it working satisfactorily again!
Has anyone gone any data that shows the frequency of old PACS films being reviewed. e.g. on average a film is reviewed twice in the first year, 0.6 times in the next 5 years 0.02 times in the next five years after that etc?
Infinitt RIS is designed to incorporate voice recognition (Speech Recognition) engine which can entirely eliminate transcription and report approval. Moreover, Infinitt RIS can provide the separate transcription functionality which is comprehensive and flexible Microsoft word and it can be tightly integrated into the system. User-defined templates are designed in accordance with center's workflows if needed. The completed report is distributed to the referring physician through the printed/faxed/or e-mail or put into the batch for the later distribution after electronic signature.