Medo is the platform which acts as an additional hand for radiologists by assisting them in diagnosing any possibilities of a medical condition by reading an ultrasound scan. Apart from AI predictions it also does few basic activities which a normal PACS (Picture Archiving and communication system) does like viewing DICOM images, basic editing, annotating, report preparation, image comparison and collaboration with teammates.
I had a limited knowledge about imaging technologies like Ultrasound or MRI. I was curious about the world of Radiologists, their day to day life and challenges faced by them. It was a little intimidating but exciting at the same time. Our client helped us in setting up a few quick meetings with their Radiologists friends in Pune at their labs. The information we got from these meetings was immensely useful and helped us throughout the project.
The major goals of the system were:
Our visit to the Radiology lab helped us in observing their work environment closely. We spent half a day there in Pune in their ambient light office. We had a detailed walkthrough of the process of how radiologists browse and annotate multiple images in a scan and prepare a report. By the end of this meeting, we had a decent knowledge about radiologist’s workflow and PACS system
Given the time of the project, we had to iterate on multiple flows and get quick feedback from all the stakeholders. We had few iterations of wireframe prototypes, where we quickly iterated on core user flows using Invision. Early feedback helped us in driving the design into the right direction. Having multiple radiologists in our stakeholder team was a real asset for us.
Worklists are a way for a users to manage multiple patients' scans in groups. Worklists help users in staying focused on their work and hide all clutter. Users can make their custom worklists based on procedures they are expert in and scan priority.
We broke down the upload scan process in 3 simple steps. Each scan must be assigned to the patient at the time of upload itself. We covered the scenario for a new patient and returning patient.
We broke down all possible functionalities which can be performed on any scan into 5 categories. View/edit scan, AI result, reporting, seek opinion and patient documents. We had to make sure that radiologists should be able to scan through multiple images in very less time. Along with basic functionality which most PACS system have, where users can scroll on image to traverse through images on any specific plane (x, y or z) we also designed a gallery viewer. Gallery viewer can be expanded using a little toggle button at right, which also increases the thumbnail size.
AI engine designed by the team can go through a scan as a whole (all images) to predict the result, however, based on our discussion with stakeholders, we all agreed that radiologists should be able to choose specific images to get AI result. If they agree, they should be able to add findings to a user report.
There are lots of interesting things happening all over Medo’s interface, like AI results, collaboration with colleagues and comparing scans but the final outcome is a simple report, which will be printed and handed over to the patient or email to Doctor. Reports have their own template based on the scan type.
Comparison has immense value for radiologists, it makes their life much easier. They don’t have to look into multiple folders, plastic slides or printed papers to compare patient reports, in order to access health status.
I had done a few projects related to physician appointment booking, queue management and others but Medo was much different and more complex project for me, where I got a chance to have a much deeper look into the Imaging world of healthcare industry. Doctor’s time and diagnosis accuracy have a paramount value in healthcare, this was the biggest learning for me. Little time wasted with fiddling interfaces can delay the patient’s report, which further can have a very unpleasant outcome.