The purpose of this chapter is to undertake a thorough study of the problem of varia- tions in ultrasound reporting and explore some of the possible solutions to this prob- lem. First, we study the process of an ultrasound examination in order to understand the importance of reports in ultrasound. We understood that because ultrasound exam- ination involves at least two different personnel, the clarity of the reports produced is very important as they serve as a communication tool between the two personnel.
Following that, we explore the problems of variations in ultrasound reports where it can be divided into two major areas which are the reporting styles and format and the terminologies used. Ultrasound reports can be written in two different forms which are the free-form and the structured form.
Cramer et al. [18] further classified the structured form into three tiers. The first tier is where reports have contents grouped under a specific heading. The second tier on the other hand improves the first tier by having the content itemised. Finally the third tier further improves this by standardising the language used.
We learned from this that the 100 sample reports we collected belongs to the first tier of structured reporting. However, we observed that the first tier structured reports were not enough to reduce the problem of variations in ultrasound reports. From this,
we decided that the medical ultrasound reporting system developed in this research should comply with Cramer et al.’s third tier of structured reporting.
To make sure that structured reporting is the best option to solve the problem of variations in ultrasound reporting, we then evaluate the potential benefits that struc- tured reporting could bring. This leads us to five benefits of structured reporting which are the production of a standardised report, consistent terminology usage, improved accuracy of the reports, increased accessibility of data and higher reproducibility of the reports. We found that these benefits complied with Ellis and Srigley [23] seven key parameters of quality reporting namely, timeliness, accuracy, completeness, con- ferment with current organisation standard, consistency and clarity in communication. Although we found many benefits of structured reporting, there are several reasons why wider implementation of it fails. Review of several literatures lead us to three main reasons which are the rigidity of the report, technological constraints and the radiolo- gist’s resistance to change. Out of this three challenges, the radiologist’s resistance to change is seen as the biggest obstacle in implementing structured reporting.
Following this, we look at the possibility of our research improving this. First, the medical ultrasound reporting system that we developed should be able to standardise terminologies used in the reports. This has led us to the usage of ontology because of its ability to share common knowledge and enables reuse.
Our study found that there are several existing biomedical ontologies that can be used in our reporting system. However, these ontologies were too big to be imple- mented in a specific application. Following this, we decided to reuse classes from existing ontologies to develop an ontology that is more specific to our domain.
Finally, we investigated the possibility of giving flexibility to the radiologists in us- ing the reporting system by allowing them to write their reports in free-form which will later be transformed to structured form. We found RST as the best mechanism to do the transformation because of its ability to recognise and retain relations between text spans. As a result of this, we investigated previous works that have implemented RST in discourse parsing whilst putting emphasis in the possibility of combining ontology together with RST.
The study in this chapter brings us to the conclusion that it is possible to reduce the problems of variation in ultrasound reporting by standardising the reporting style, format and terminologies used. It highlighted that the terminologies can be standard- ised using an ontology while the reporting style and format can be standardised using structured reports generated by RST.
The Design of a New Ultrasound
Reporting System
3.1
Introduction
Humans are known to sometimes being incapable of accepting changes and adapting to new things. For an ultrasound reporting system to be accepted by the radiologists, it is important that the system has a level of flexibility that would allow for reports to be produced in a structured form without causing too much inconveniences and frus- trations to the radiologists. Studies developed by Bosmans et al. [13] and Danton [19] concluded that radiologists were found to have problems by not having more options in creating their report. This chapter presents a system architecture model that allows for these radiologists to have more options and flexibility in creating their report. It ex- plains how the overall system works as well as how each component plays their part. The focal point of the system architecture model is the structured report page. This chapter will discuss in detail about the design of the page including the guidelines that were used as references. The overview of the system architecture model presented in this chapter has been published as a position paper in the 8th International Conference of Health Informatics [94]. The system architecture model was also presented as a poster in the 2015 UK Radiological Congress (UKRC) where a survey on the ultra- sound reporting system was also carried out. This chapter presents the result of this survey and discusses the issues that were raised.