Friday, March 30, 2012

The Dilemma of Data

Logging data is one of those things that every single person diagnosed with Type 1 diabetes probably does initially, but that falls by the wayside after the first few years (or months!). It's one of those things that we all know helps, yet most of us struggle to do. We are told on one hand that diabetes doesn't have to be something that rules our lives; yet, it's hard to frame something as just a daily nuisance when it involves not only so much daily minutiae but also involves recording every single instance of minutiae into a paper logbook or computer system.

I personally find logging data manually to be incredibly useful. Sure, my pump records a lot of information, but there are things it doesn't record (like specifically what I ate) that I might want to know. Plus, the act of manually recording blood sugars and other events makes me much more aware of what is going on than when I don't record. For me, as someone with a visual impairment, the advent of computer software has also greatly increased the accessibility of keeping a logbook. Despite this, I still struggle with exactly why I'm taking the time to record all this information.

A paper logbook and iPhone diabetes management app Like other aspects of diabetes over the past few decades, logging has changed a lot. Those of us diagnosed before computers were common used ruled notebooks or binders full of loose leaf paper. There was often no data analysis component—we logged so that we could show it to our doctors, because meters back then did not have memory. In the late '90s meters began to come out with computer connectivity, and it seems to be at this point the whole idea of "data analysis" developed. These days some people still use pen and paper, but people also have the choice of using a myriad of computer applications, websites, apps, and computerized meters and pumps to collect, store, and analyze relevant data.

For me, I don't actually care that much about blood sugar statistics. Sure, it's great for tracking progress, but it doesn't actually help me improve my control in any way. Plus, almost all meters and pumps these days are able to provide some basic information about averages and glucose ranges, either through the meter itself or through accompanying software. So, when software is expecting me to enter data manually—which, as I said above, I don't actually mind—I really wish that it would provide me with actual tools I could use to improve my control. The current software out there seems little better than a paper logbook in providing these types of tools.

What I really want to see are causes and effects. I don't care about the fact that my average blood glucose is 8.2 mmol/L or that I've been waking up low every morning for the past few days or that I've tended to go high around dinnertime lately. I'm aware enough that I have a sense of this information without the nuisance of logging everything—and the very process of manually logging makes me even more aware of these trends. What I am interested in seeing is the cause and effects of variables on blood sugar, so that I could either control these variables or adjust my insulin to get tighter control.

With the huge craze on being able to quickly collect, skim, and analyze various types of data out there, diabetes software seems to be lagging way behind. Designers seem stuck on this blood-sugar-statistics model and on the before-and-after-meal framework for organizing data. What I would really like to see is diabetes software that views blood sugar as a dependent variable, and allows tracking of some standard independent variables (food, insulin, exercise) as well as the creation of any other independent variable event one wishes to track. Plus the ability to create parameters for these variables (time, intensity, or anything else) and filter results based on these parameters. Plus the ability to then filter blood sugar in various time relationships to one or more of these variables (before, after, during, between). Plus the ability to tag readings or events with any relevant notes and then search through and organize the tags at a later date. And then provides tools for visualizing and analyzing blood sugar in relation to all these other variables that have been collected. Want to see whether blood sugar always rises on the third day of an infusion set? Want to see how much exercise of X intensity lowers blood sugar compared to exercise of Y intensity? Want to see whether exam stress really affects your blood sugar levels? Want to see if you really run higher during the winter compared to the summer? Want to see if you go low every time you do X even though it's something you only do once a month? Want to see if an hour of swimming lowers your blood sugar differently than an hour of cycling? It's little things like this that would make manually recording information truly valuable.

As of yet, there is no software that even comes close to what I am describing. I am currently using a combination of a desktop application and an iPhone app that are okay, but far from perfect. And I think this, more than anything, is why I really struggle to log sometimes—because I feel like the data I log is never really taken advantage of and used in a truly meaningful way that will let me figure out some of the factors that make managing Type 1 diabetes so maddeningly "random" sometimes. Isn't diabetes really about problem-solving in the end? I want software that will actually help me sort out data and make decisions about how to solve problems, not just tell me statistics about my blood sugar.

No comments:

Post a Comment