Radiology

Dr Graham Hurst Director of Radiology at Norfolk and Norwich Hospital has produced a case study of his experiences, presented at HC 2003 in Harrogate.

 The conclusions speak for themselves:

 1. The GHG Radiology Solution dramatically reduces the time taken between the radiological examination and the issue of the report.

 2.  Voice recognition reporting is at least as fast for the radiologist as a conventional tape dictation system, once full training on the system has occurred,

 3. Fewer errors are made compared to a secretarial transcription system

 4.  Highly cost effective due to the reduced clerical time required to complete a report.

The full report:

Efficient report generation and distribution in Radiology

The Radiology Department of Norfolk and Norwich University Hospital generates approximately 110,000 radiology reports annually. These reports provide essential information to general practitioners and hospital doctors about the patients’ x-ray tests. A prompt delivery is essential.

Conventionally radiologists dictated directly to a secretary who would transcribe the document on a typewriter. This practice has largely died. It was replaced by dictation into a tape/handset.This is associated with a variety of problems not least delays in transcription onto the radiology information system (RIS) but also difficulty understanding the tape and loss of tapes periodically.

Dr. Graham Hurst, consultant radiologist, began to look for an alternative method of report generation. His goals were to increase the efficiency of the department, improve the quality of reports generated and reduce administrative costs. Voice recognition seemed a likely candidate for investigation.
 
Dr. Hurst worked closely with GHG's Software Developments of Milton Keynes to build a voice recognition system carefully customised for the hospital. GHG integrated their own VoicePro command software with ScanSoft's Dragon NaturallySpeaking® 7 speech recognition firstly to enable the control of the McKesson TcRad v 8.9 RIS with Voice Commands and secondly to dictate radiology reports on to the RIS for distribution around the hospital and general practitioners.
 
Keystone Roamer software from Words Worldwide was employed to enable voice profiles to be accessed from any of the nine networked PCs. Following thorough training delivered by GHG Software Developments six radiologists now regularly use the system and more are in the process of being trained. A customised radiology vocabulary has been developed by the Radiology Department and GHG to enhance the recognition of specialist radiological terms.
 
The radiologist works in front of a PACS radiology workstation in which the electronic images are displayed. Adjacent to this is the RIS computer with voice recognition functioning. The radiologist talks the attendance number for the examination into the RIS computer and this activator system brings up the reporting screen for that patient. The report is entered and checked. On the command "authorise and print" the system distributes the report electronically, prints a paper copy and then automatically brings up the entry screen for the next patient examination attendance number to be dictated.
 
Each time a dictation session is completed the user runs the Vocabulary Builder feature in Dragon NaturallySpeaking to add new words and terms used to further improve accuracy. Dictation shortcuts or ‘macros’ are used when appropriate. These are a feature of Dragon NaturallySpeaking and are mostly used for reports on examinations with normal results.
 
Initial results of the system are encouraging.
 
The speed of delivering reports appears to be faster which means that referring doctors are receiving reports as soon as the report is created.
 
28% of reports (30,000 per year) are now generated with voice recognition.
 
Users also reported positively on the effect of using voice recognition;
 
"It is satisfying getting reports promptly to the rest of the hospital",
 
 and also commented on the improvement in workflow for the hospital as there was:
 
"no checking of secretarial work later in the day/week"…

Following an internal analysis of the overall effectiveness of the system Dr. Hurst sums up the
benefits in terms of efficiency, accuracy and cost effectiveness:
 
“A full evaluation of the system by Dr. Hurst and colleague Dr. Sameen Rana is planned in conjunction with the University of East Anglia but all signs point to voice recognition as a valuable asset for the future of efficient procedures in radiology”.

Contact details:

Dr. Graham Hurst
Norfolk and Norwich Hospital NHS Trust
E-mail: graham.hurst@nnuh.nhs.uk
Tel: 01603 286104