Novo Nordisk
Healthcare
Challenge
Until now, the only way for the clinical team to ensure patient safety and the effectiveness of a new drug was to identify deviations and anomalies in an Excel file containing thousands of lines.
Identifying discrepancies and outliers is particularly difficult, and errors can easily occur.
Solution
I transformed the experience of the clinical team by designing a seamless, visual and collaborative one-stop-shop for all clinical data review, allowing Data Managers and Medical Reviewers to perform daily checks with improved accuracy.
User research
Ecosystem mapping
Concept Development
Visualization
AI
UI & Prototyping
Usability Testing
Accessibility
User research
With the help of a Strategic Designer, I began by mapping out the existing ecosystem and immersing myself in the checks and software that scientists use on a daily basis.
Thanks to my visualization work, my product team was able to look beyond the technical complexity to envision real user journeys and identify their pain points.
Iterations
I then conducted an extensive iteration phase to thoroughly rethink the structure and hierarchy of the interface, evaluating the impact and relevance of menus, charts, spreadsheets, tabs, and panels on the readability and comprehension of the information.
Combined with initial user testing, this allowed me to define a key concept in which the visual elements play a central role in analyzing trends, and is linked to data points so that no essential information would be overlooked.
First MVP
Working closely with a front-end developer, we developed a React prototype in 3 months to test our hypotheses and convince management to begin the transition to a new interface.
It was finally decided to develop it using PowerBI, and while the development team handled its implementation, I focused my efforts on two specific areas: adding AI features and accessibility.
Empowered by AI
To help users identify special cases, our team of data scientists built an AI to auto-flag deviating points and recommend text for queries.
My role was to think about the best way to visually highlight the content created by AI to increase efficiency and build trust in the data.
Improving accessibility
Accessibility isn't an option, it's a principle of equality that benefits everyone. That's why I made a point of improving the readability of my MVP, particularly by redesigning the graphics, so that no information would be misinterpreted.
Using an eye-tracking system, I was able to measure the improvement between the old and new versions among about twenty users: the users were three times more efficient and no longer made any mistakes.
It made me realise how much accessibility testing is a really powerful tool for decision-making.
Result
After a year of work, this solution was implemented via a Power Platform, and we were able to demonstrate clear interest from users thanks to a number of metrics, such as the number and frequency of tasks performed via the interface, as well as the growing diversity of profiles wishing to use this product.
I also equipped Novo Nordisk's Design System with a set of accessible data visualization components that serve as a guide for most internal professional tools.
Clinical trials
Total users
Daily users
Potential new users
time saved performing daily checks
satisfaction (compared to 32% with the previous software)
