Data and what to make of it

2 11 2009

I just downloaded data from my Dexcom and have been taking a look at it.  There’s so much data and so many different ways of looking at it; it’s hard to tell what to make of it.  There are a few items I like to focus on:

Modal Day screen

  • average blood sugar for the month
  • average blood sugar for the past 3 months
  • standard deviation

Glucose Distribution screen

  • % in range for the month
  • % in range for the past 3 months

Success Report screen

  • compare data montly
  • compare data quarterly

The average will tell me about what my A1c will be.  I use this chart and I have found that comparing my Dexcom average to this chart is very close to my actual A1c.

The standard deviation will tell me if I am doing too much of a rollercoaster.  Lower is better.  I will confess, mine is not as low as I would like, so I know that I need to level it out.

% in range is very important to me.  Knowing that I am in range 75% of the time is greatly empowering.  Knowing that I am 95% in range upon waking is even more empowering.  Of course, knowing I am only 50ish% in range after lunch tells me that I need to work on that area.

Comparing the data from month to month is great for trends.  I can see that my average in October was less that what it was in September, which is great.  I can also see that my average for the last quarter is lower than the previous quarter, so I imagine that my A1c will be lower as well.

Using the Dexcom software can be a little overwhelming (there’s so much more data available than I even mentioned), but if I focus on these few things, I feel like I have a pretty good handle on my diabetes control.


Wavesense vs. Freestyle data

14 05 2008

I’ve been using the Keynote for about a week now, so today I sorted through all the data and made a few graphs and analyses. (Let me know if you cannot view the link)

meter comparison spreadsheet

Here are some of the observations I saw from the data:

  1. The overall average is nearly identical. The Freestyle (for the first few days of the spreadsheet, I was using the Flash, then I switched over to the Lite. I don’t think it made much of a difference) had an average of 124.9, while the Keynote had an average of 125. UPDATE: I removed a couple lines where I only had a number for one of the meters (but not both). This resulted in a Freestyle average of 124.5 and a Keynote average of 123.5.  Still not a huge difference, but in this case, the Keynote gives a LOWER average.
  2. The Keynote has a lower standard deviation. Freestyle’s standard deviation was 57.4, while the Keynote has only 50.5. This is very important. Lower standard deviation means fewer/less extreme values. This is what they mean by the “plus or minus 20%” or whatever. A lower standard deviation means it can be “plus or minus” by a lower amount. This indicates increased accuracy, which is what Agamatrix claims in the first place. As a result of my update in #1, the Freestyle std dev is now 57.7 and the Keynote is 49.6
  3. Related to #2, the Keynote produced a higher average when blood sugar was 100 or below; and a lower average when blood sugar was above 100. Again, this indicates a lower standard deviation and fewer extreme values.
  4. The biggest difference in standard deviation was when blood sugar was above 100. I feel that this is very important, because that means high numbers are more accurate. As such, there is lower risk of over treating a high and thereby ending up low.

If anyone wants to offer any more observations they find in the data, feel free to leave a comment below. Try to keep the comments about the differences in Freestyle vs. Wavesense and not about my numbers/care/treatment/etc.