A mockup of a person checking out the Social Change Innovators website on a laptop

Social Change Innovators

SCI is a knowledge base and online learning environment for social entrepreneurs. I co-created the platform together with Dr. Debbi Brock, redesigned it multiple times, and added a learning management system.
Design
Frontend
Backend
Client
Catalyst 2030, Social Enterprise Netherlands
Years
2018 - 2024
Role
Lead Developer

XCEPT

This project is a collaborative initiative between the University Hospital in Groningen and XVIVO, focusing on enhancing liver transplant outcomes through advanced data analytics.
Design
Frontend
Backend
AI
Client
University Medical Center Gronignen (UMCG), XCEPT
Years
2020 - 2022
Role
Designer, Developer
A mockup of the web application XCEPT used inside a hospital

Technologies

Python Backend
MySQL Database
React Frontend
Raspberry Pi
About the project

A decision support system for normothermic liver perfusions

During my master’s thesis in Artificial Intelligence, I initiated the development of a system that integrates data from various medical devices used during organ perfusions. The system facilitates advanced data analysis to aid surgical teams in making data-driven decisions, thereby increasing the success rates of liver transplants. Following the thesis, I have continued to refine the system, working as a developer and designer to adapt it for use across multiple medical centers.

Why?

The global increase in patients on transplant waiting lists demands innovative solutions to expand the number of viable transplants. One method, machine perfusion, simulates body conditions to preserve organs before transplantation. Surgeons rely on data from these perfusions, generated by diverse devices such as blood gas analyzers, to assess organ viability and necessary interventions.

Main Problems

Lack of integration
Currently, there is no unified technological system in organ perfusion units. Multiple devices operate independently, providing disparate data outputs.
Inefficient Data Management
Much of the critical data is still recorded manually on paper, making real-time data analysis and decision-making challenging.

Process

Research

Observations

Expert Interviews

Process Mapping

A process flow chart of normothermic liver perfusions
The process of one type of normothermic machine perfusions

Prototypes

In an iterative process, I designed an initial interface. With a very short feedback loop to endusers, the designs were optimized.

Initial sketches of the XCEPT interface
The layouts and components on of the phase specific pages
Initial sketches of the XCEPT interface

Development

After the design phase, the web application was developed with a React (Typescript) Frontend, Flask (Python) backend, and MySQL database.

A screenshot of XCEPT

Features

An illustration of the internal network of XCEPT that allows external users to access the web interface

🔒 A secure local network

Multiple devices sent live data to the Python server. Viewing the web application from outside the local network is possible through a secure proxy server.

👀 Computer Vision

Not all devices used in the process supported data export during the procedure. Therefore, we attached Raspberry Pi's with cameras to the machine and used digit recognition to generate a live feed of perfusion data. The data is sent though a local network to the server.

An illustration how computer vision is used to read the display of a medical device

✨ AI Recommendations

At any given time, the system analyzes the current values and recommends potential medications that could improve the liver viability. The training data are past actions by surgens in reaction to bloodgas and electrolyte values.