Knowledge Hub.

Exeter Diabetes App – Raising Clinical Awareness.

Identifying patients with MODY can be challenging though, so tools are needed to help doctors work out which of their patients are most likely to have the condition. Researchers at the University of Exeter Medical School and Royal Devon and Exeter NHS Foundation Trust developed statistical models to assess the likelihood of a patient having MODY based on their clinical features such as their age, BMI and treatment.

OCC were approached by these researchers to help turn these statistical models into an online calculator. This tool, the Exeter Diabetes App, could also raise awareness by providing information on the condition and on other forms of diabetes. The team’s vision was for an app that would be generally available to both medical professionals and patients, so a clear and intuitive user interface design was a high priority.

Making an App for Everyone

OCC designed a simple home page layout with several options. Based on their selection, the user would be guided to either a calculator input screen for MODY or to an external website with additional information about the condition. On the calculator, the user enters a set of data inputs as prompted by the app which are validated to ensure their correctness. This information is passed to a clinically validated algorithm developed by University of Exeter researchers which calculates the probability of the user having MODY. The outcome is shown on a results page along with additional information and advice which is tailored for the patient.

Oxford-Computer-Consultants-Exeter-Diabetes-App

Oxford-Computer-Consultants-Exeter-Diabetes-App-Results

To maximise the possible usage of the app, the customer requested that it work with all major browsers and on phones and tablets as well as PCs. In addition, because clinicians sometimes work in locations without readily available internet access, the app should also be downloadable and usable offline. To meet both requirements, we used a relatively new technology – Progressive Web Apps (PWA). This allows developers to create an application which can be accessed as a website but which can also be downloaded like an ordinary app on phones and tablets. Taking this approach allowed OCC to create a multi-platform application without the challenges of adding it to the Apple and Android app stores.

The base of the application is React using TypeScript. By using create-react-app, OCC could build the app from a clear starting point to become a progressive web application, enabling us to provide the offline requirement necessary. We incorporated Bootstrap for elements used within the application, which allowed for a clean modern look, greater accessibility plus faster development time.

Developing Software as a Medical Device

The Exeter Diabetes App falls within the scope of EU Medical Device Regulations and so is classed as Software as a Medical Device (SaMD). This classification requires a very high level of documentation and testing as supporting evidence of its safety for use in clinical situations. OCC have considerable previous experience in delivering medically approved software and mHealth apps, which supported our design and engineering approaches for this project. We worked in partnership with Beverley Scott from BCS Clinical Consulting to ensure that our documentation and approach met MHRA standards. This collaboration has been successful working on the DECODE project also.

SaMD standards require the implementation of a rigorous and traceable testing regime. The JEST and Cypress frameworks were used to assist this aspect of the development. JEST was used for traditional unit tests, ensuring that the outputs of the algorithms used in the calculator were correct. Cypress was used for end-to-end (e2e) testing in which the actions of a user are simulated to ensure that all the functionality performs exactly as it should.

Future Work

Currently the app only allows for the calculation of MODY, but researchers at Exeter have also been developing statistical models to aid classification of more common forms of diabetes and it is hoped that a future version of the app will include a further calculator for Type 1 and Type 2 diabetes.