Doctors verify each diagnosis, improving the machine learning engine
Enhancing the quality of each patient's face time with their doctor
Our main task was to deliver a machine learning-based app, designed to improve and shorten the appointment time with a doctor, thanks to a preliminary analysis of symptoms provided by the patient.
Patients answer smart questions and boost diagnosis accuracy
Our team analysed relevant focus group studies as well as marketing segmentation data. We put the clients hypotheses through our verification process, and created a sitemap with user flows pointing out the strenghts, weaknesses and opportunities.
Then we hosted a workshop with the stakeholders and summed up the business, techonology and qualitative data we were basing our recommendations on. During this workshop, we created personas, user scenarios and user stories, which led us to identifying the scope of the MVP.
Scenarios from the workshop were divided into short stories that were shared with the whole team and used for progress tracking and backlog creation.
We kicked off our design sprint with ‘crazy 8's’ to generate initial ideas. Once these had become sketches everyone was clear on, the process moved on to wireframes and then visual design.
We started by developing a style-guide, so that the developers had a consistant and flexible framework that remained in-line with the brand.
Bingli back-end architecture makes use of Microservises. The project also utilises Amazon ECS, providing robust infrastructure for the machine learning engine.
During the patient's interview, questions are adjusted in reponse to previous information and answers provided.
The bayesian network within the machine learning engine, factors in the prevelance of diseases alongside the symptoms provided.
This computation of both probablity and the symptoms allows the system to generate a highly reasoned suggested diagnosis for each patient.