ICTD Capstone Presentations (CSE 482B). Tuesday, June 6, 2:30-4:00 pm, CSE2 371 The ICTD Capstone (CSE 482B) brings together groups of students to work on projects to provide solutions for low resource environments. Based on an initial project idea, the students are charged with designing and developing a prototype system . This quarter, the projects were drawn from the domain of Global Health. Two of the projects were motivated by the availability of low-cost mobile EKG devices which can help diagnosis various heart diseases, such as atrial-fibrillation. The projects consider aspects of how these devices could be used in low resource settings to triage heart conditions. The other two projects were in collaboration with doctors from Children's Hospital who are developing new protocols for detecting respiratory problems in Uganda. They have developed a mobile application (ALRITE: Acute Lower Respiratory Illness Treatment and Evaluation) for supporting a diagnostic protocol, and the students extended the system to support scalability to large deployments. Students will give final presentations and project demos on Tuesday, June 6, 2:30-4:00 pm in CSE2 371. The presentations will also be available on zoom: https://washington.zoom.us/j/98302897925 The projects are: Integrating mobile apps with OpenMRS The current ALRITE application saves patient records locally to the device which limits opportunities for collaboration and scalability among different clinics in Uganda. The project integrates the ALRITE Application with OpenMRS, an open-source medical record system to create a streamlined syncing process for patient records. The integration between ALRITE and OpenMRS enables creation of patient records, recording visit information and the diagnoses and treatment plans generated by the mobile application. This project is a proof of concept that shows that OpenMRS is an appropriate backend for medical protocol apps, and gives a pathway for deployment of these apps in multiple domains. A Platform for Customizing Protocol Apps The ALRITE application is a stand alone Android Application. This means that whenever changes are needed to update wording, data fields, or the branching logic, it is necessary to go back to a software developer, a time consuming and expensive proposition. The project was to develop a digital editor which would allow a health researcher with limited coding knowledge to edit the existing ALRITE application to create a new diagnosis workflow. This workflow is then sent to pre-existing question templates which are filled in with the data received from the editor, resulting in a question by question examination ultimately leading to a diagnosis. Templates include text input, multiple selection, and multiple choice, the digital editor serves as a webpage which allows for recovery of previous workflows and easy editing of existing workflows. Explainable AI for Interpreting ECGs Currently, there are many ML models that, given a patient's ECG, can diagnose whether that patient has A-fib with high accuracy. The issue with many of these models is that they are black-box models which make it difficult for community healthcare workers to interpret their decisions. Our solution is AFib-XAI, a program that uses explainable AI methods to display the important data points and regions of interest in an ECG that were factored into a model's A-fib diagnosis, as well as generate accurate, human-readable explanations for that model's diagnosis. AFib-XAI can be run through a command-line interface. The program offers a selection of 3 A-fib diagnostic deep learning models and 4 SHAP-based explainability methods. The user must provide an ECG, make a model selection, and select an explainability method. Afterward, the program will be run. Once the program has finished running, the user will be given a visual result from the SHAP explainability method, as well as a very basic text explanation generated from the results. ECG Training Application for Community Health Workers Mobile ECG devices make it possible to perform community screening for heart conditions. When these screenings are performed by front line health workers, it is necessary to provide some basic health training to the workers to help them explain the results to patients. This projected developed an ECG training app designed specifically for Indian Volunteer healthcare workers (ASHAs). The app has training lessons on the basics of reading ECGs tailored specifically for ASHAs, and includes practice questions and progress updates. Through user testing, the design and lessons are simple and easy to navigate, giving the app the potential to be utilized by ASHAs in the future.