Structured reporting as the potential future of radiology reporting
One of the main benefits of structured reporting – which are techniques for creating, storing, and distributing reports – over dictation and speech recognition is that the end product more closely resembles the shape and content of the traditional reports whereas the other tools are largely output in long strings of words, said Curtis P. Langlotz, MD, PhD, associate professor of radiology, University of Pennsylvania, during the Communicating Results session at the annual meeting of the Society for Imaging Informatics in Medicine (SIIM, formerly SCAR) in Austin, Texas, yesterday.
Also, the efficiency of speech recognition technology depends greatly on the user. Users that create a body of macros for use in the reports, for instance, are those who tend to be more efficient in using speech recognition-based tools reports. Other variables also exist such as the recognition accuracy of the system, the user interface and the user’s experience with computers, said Langlotz.
As for general benefits of structured reporting, the technology can: reduce transcription costs, offer quick distribution, reduce errors, might be faster than dictation, is preferred by some physicians, is high-quality, and often include decision-support assistance for difficult cases, said Langlotz.
The world of structured reporting isn’t perfect, however. As with speech recognition technology, these tools might provoke a radiologist to needlessly watch the report as it is being generated instead of focusing on images, Langlotz said.
To improve communication of results overall, Langlotz suggested that consistent imaging terminology (or lexicons) are needed. The RadLex project supported by RSNA is as an effort to remedy this situation as it attempts to provide a comprehensive universal language for radiology. So far the group has developed 5,308 anatomic concepts. It is a set of terms to associate with clinical reports in an EMR, and is a successor to the ACR Index for retrieving teaching files, as well as a good data collection tool, he said.
Adoption of a universal vocabulary for radiology will certainly aid the use of structured reporting which Langlotz believes will be a part of all reporting in the future, at least to some degree. Future systems will hopefully be as fast as conventional reporting used today but also involve image displays and some aspects of speech recognition as well, he said.
Also, the efficiency of speech recognition technology depends greatly on the user. Users that create a body of macros for use in the reports, for instance, are those who tend to be more efficient in using speech recognition-based tools reports. Other variables also exist such as the recognition accuracy of the system, the user interface and the user’s experience with computers, said Langlotz.
As for general benefits of structured reporting, the technology can: reduce transcription costs, offer quick distribution, reduce errors, might be faster than dictation, is preferred by some physicians, is high-quality, and often include decision-support assistance for difficult cases, said Langlotz.
The world of structured reporting isn’t perfect, however. As with speech recognition technology, these tools might provoke a radiologist to needlessly watch the report as it is being generated instead of focusing on images, Langlotz said.
To improve communication of results overall, Langlotz suggested that consistent imaging terminology (or lexicons) are needed. The RadLex project supported by RSNA is as an effort to remedy this situation as it attempts to provide a comprehensive universal language for radiology. So far the group has developed 5,308 anatomic concepts. It is a set of terms to associate with clinical reports in an EMR, and is a successor to the ACR Index for retrieving teaching files, as well as a good data collection tool, he said.
Adoption of a universal vocabulary for radiology will certainly aid the use of structured reporting which Langlotz believes will be a part of all reporting in the future, at least to some degree. Future systems will hopefully be as fast as conventional reporting used today but also involve image displays and some aspects of speech recognition as well, he said.