AGIL
ACTIVE GLUCOSE INSULIN
CONTROL SYSTEM
Though we developed a design that addresses the main needs our users identified, there is still substantial work to be done on a variety of fronts before our system can become reality. Technological progress in the fields of blood glucose monitoring and insulin delivery is ongoing, and devices that integrate these capabilities in the way we envision should be available in the near future. The challenge is to equip these devices with interfaces that not only minimize errors and support simple, highly efficient interaction, but also make users feel in control of their health, and respond seamlessly to a wide variety of users and situations.
A misjudgment that we initially made, and that other PDA-based solutions seem to make, is that people with diabetes want to be frequently presented with an arsenal of information about their health. Designs that put tasks such as viewing charts or delivery histories and consulting on-line resources on the same priority as the basic tasks of delivering insulin, and checking glucose levels did not connect to the priorities of the users we consulted. These simple tasks are done several times a day, every day of the user’s life. People who are used to monitoring their blood sugar generally like the idea of having these features available, but do not need to be confronted with them every time they use the device.
We feel our design has demonstrated the effectiveness of this return to simplicity in the majority of user interactions, while still preserving the ability to perform more complex tasks. The most compelling of these interactions is the ability to view visualizations of various factors that can affect blood sugar management. The device would already be recording information about insulin deliveries, glucose levels, and indirectly, meal times, sleeping habits, and periods when the pump is removed. Collecting this information in an unobtrusive manner was a focus of our design, as all of this data is collected as a result of some action taken in the interface. Possible useful visualizations we did not design include:
· Tracking glucose levels for each schedule, to determine whether the amount of insulin is correct, or was underestimated or overestimated.
· Cross-cutting days to view only levels at waking up or going to sleep, for example.
· Methods of identifying outliers and analyzing them, as these may help aid the user in modifying potentially harmful items, such as foods that contain more carbohydrates than normal.
By using a PocketPC-based design, we were able to add color to graphs and make provide some means of manipulating them. However, a great deal of work still needs to be done on visualizing this data on such a small screen. An important step is to identify the data and the views that are important to users vs doctors. While patients are increasing knowledgeable about their own care and diabetics implement the bulk of their treatment themselves, a team of medical professionals can still utilize data in ways that users cannot. This means that including these special visualizations on the device should be strongly questioned, because we have already seen that users do not want to be burdened or confused by information they can not or do not want to use.