Insight into usability

User research on usability

August 26, 2016 – In April and May this year, 32 people came by Inreda in order to help out with a research on the usability of our artificial pancreas. The goal of that research was to find out more about how people use the device, and find out what points and aspects can be improved.

The participants were split in 2 groups based on the treatment method they currently use: one group was composed of insulin pen users, the other was made of pump users. That simple distinction could help to compare the results. All participants were asked verbally to perform 3 tasks:

  • Task 1: to pair the artificial pancreas with the heart rate belt
  • Task 2: to replace the insulin cartridge
  • Task 3: to replace the batteries

To do so, we provided them with one artificial pancreas, and its user manual.

It is worth saying that none of the participants had prior experience with our artificial pancreas, we gave no single instruction regarding how to use our device or even the way it works, and during the assignment they could not ask any questions or get support from the researcher.

How well were the tasks performed ? And how did people experience the usability of the device ? This was researched after the test with a list of questions and an interview.

In general, the tasks were performed successfully, and our artificial pancreas was experienced as user-friendly. If you look for example at the table below, both groups (pen users and pump users) were able to pair the heart rate belt with the artificial pancreas. Both groups score a high success rate :


(As you can see here, each task is actually composed of a serie of sub-tasks that the user had to find out by going through the interface. Those sub-tasks were not told to the participant beforehand).

Same was true for the replacement of the batteries: both groups performed rather equally.
However, there is one subtlety: the replacement of the insulin cartridge. It was a bit expected but insulin pump users were more successful in performing and completing the task than pen users. Pump users simply have more experience in doing so, and it appears as well that some pen users could have been confused by the words being used on the interface of the artificial pancreas.


You can see in the table above that pen users’ performance especially dropped at the last step of the task : filling in the cannula (‘canule’ is the dutch for ‘cannula’) and making it dripping to ensure there is no air trapped in the whole length of the infusion set.

For those still wondering, the cannula is the ‘soft needle’ which is inserted through the skin to deliver the insulin or glucagon in the body:

We believe the artificial pancreas should be accessible and usable by everyone (not only current pump users) so we definitely will pay attention to the wording and will look for ways to make the instructions even clearer. For example with illustrations.

We also gathered an amount of points of which tasks and sub-tasks can be improved. We are taking this input into account already for the newest version of the interface, which will be used during the upcoming clinical trial.

You can freely read the whole study by clicking on that link:
Right after taking part in the study, some participants did share their experience about our artificial pancreas on Facebook. You can read the original thread by clicking here.

We want to thank everyone, who came from all over The Netherlands (from Limburg to Groningen!) in order to help us out evaluating the interface of our artificial pancreas. You really contributed to make our device better and more user-friendly. We will take all your input and implement it in the future interface. We want to thank also Rémi Meeringa, from the University of Twente, who successfully lead this research (she got a 9 for her research). Congratulations!

< Back to the news overview