Evaluation
MoSCoW List Achievement
Individual Contribution
CRITICAL EVALUATION
User Experience
During the development of our web application, we ensured that our interface was very clear and easy to use. This was done through regular meetings with our client, and ensuring the following:
- Clear labels next to text boxes
- Clearly labelled links on the navigation bar
- Clearly labelled buttons
We wanted to ensure that the interface was also nice and neat, however we had little experience with frontend development. Therefore, we made extensive use of Bootstrap for building our web application’s responsive design.
Functionality
We consistently focused on ensuring the core functionalities of the application were met; the key requirements set by our client from the beginning of the project.
We fulfilled most of these requirements, with fully-built NLP algorithms which generate summarised electronic medical records from consultation audio uploaded by a doctor. This information is also clearly accessible and downloadable, and stored in the database.
Hence, doctors are able to upload audio for creating a clinical document, and then retrieve it on the web application in order to share it with other clinicians if need be; this was a key requirement voiced frequently by our client.
Stability / Efficiency
Our project has great performance across most areas of the application, except one. For particularly long audio files, our NLP algorithms can take a long time in creating the summarised Electronic Medical Record (EMR). This is because we didn’t have time for optimising the performance of our machine learning algorithms, and also didn’t have a lot of sample audio data for testing and improving the performance of these NLP algorithms.
Compatibility
Our deployed application was tested on multiple browsers such as Google Chrome and Safari, and behaves the same without errors during core processes on the web application.
Maintainability
We believe the code of our project is well organised. We made sure that we had an easy to understand project structure, with a clear separation of concerns. Files and packages are clearly named such that their purposes are self-explanatory from their names. Moreover, features have been developed individually in their own files, functions or classes.
Moreover, an important pattern of design was separating data-driven operations (fetching and storing data in the database) from the code that renders the html pages. The main benefit of this was supporting test-driven development in the future; thus also making it easier to debug the project.
Overall, our project is well organised and helps to increase the flexibility and scalability of our application.
Project Management
Our project was well managed with responsibilities divided between with us, along with frequent meetings with our client.
However our mutually exclusive weaknesses and lack of knowledge in areas made it difficult to work together on features, and so led to a number of delays in our project. Reflecting on this, we should have improved our communication throughout the project particularly during the planning stage of the application.
Despite this, overall we still communicated well and were open to overlapping responsibilities in order to meet the client’s requirements. In addition, we would have a meeting every week which became several every week towards the end of our project.
Future Works
1. More features should be added to the web application. Users should be able to change their password and patients should be allowed to report problems in their personal information. Doctors should be able to edit the electronic medical files and store the correct record into the database.
2. The UI Design of our web application can be further improved including clearer indication of system functions and online preview of electronic medical reports. UCL and NHS icons should be added to application Home as a representative of the developers.
3. An integrated medical and clinical dictionary could be added to the system for patients to better understand the context of the electronic report and improve the service quality.
4. There are many limitations of the current automatic text summarization system. Doctors should be allowed to adjust the compression rate and the doctor edited summary could be collected as sample data for improving the text summarization system. Further research is also required to improve the overall system performance.
More testing should be conducted to find hidden bugs and complete API support should be added to the software system. Security measures is also required on the database since the privacy of patient’s personal information is of top priority of an online medical platform.