When building a new information system, the supported data values are really building blocks. They support the system itself, and they determine what the system is capable of doing.

For people with type 1 diabetes, I believe there are four different types of data needed in the system. Because I'm focused on the needs of type 1 diabetes I'm assuming the primary treatment is Insulin. But even with type 1 diabetes, there are a number of other medications that may be worth tracking. This might form another data type, though it should be considered as part of a Personal Health Record (PHR) or Electronic Health Record (EHR).

1. Blood glucose readings.
2. Insulin taken.
3. Carbohydrates consumed (also protein, fat, fiber)
   
4. Comments: sickness; exercise; stress.

The first three choices are obvious, but I also believe it's critical to be able to represent comments as a critical piece of data that should be presented in a structured way so it can be interpreted by software.

For example, if I'm using a continuous glucose meter whose readings are affected by other medications, then I may want to enter a comment with an associated range of time to indicate that these readings aren't correct.

The system must be flexible and extensible to support future requirements. The way in which data is represented should reflect that.