How do I get the best use of my Dexcom Clarity data?

Diabetes Technology Glucose Monitoring
1

I usually log in to my Dexcom Clarity account (desktop version) once a week and look back at the last 14 days.

I check my:
TIR (time in range)
SD (standard deviation).
Average glucose
Mean glucose
CV (Coefficient of variation)

I know how to interpret and get use of these parameters, but lately some new terminology has appeared.

Under statistics → hourly → advanced I now find IQR, IQ std. dev. and average SD.

I get that IQR are SD taking account for outliers (simply put).

To make this thread useful for everyone, let´s use the 70-180 mg/dl range, 70% target for TIR and a CV at 33 % or lower.

Three questions:

  1. What number should IQR ideally be?

  2. What is IQ std. dev. and what should this number ideally be?

  3. What is average SD and what should this number ideally be?

Thanks in advance for any feedback and answers on this topic.

1 Like
2

I love a good Statistics Sunday.

Lets start with definitions. https://productstore.clarity.dexcom.com/Documentation/en/Dexcom_Clarity_User_Guide_Home_User.pdf

  • Min: Minimum is the lowest CGM reading.
  • Mean: The average of CGM readings.
  • Max: Maximum is the highest CGM reading.
  • Quartile 25: The CGM reading where 25% of all readings fell at or below this value, after ranking all values from lowest to highest.
  • Median: The midpoint of all CGM readings where half the readings fall above.
  • Quartile 75: The CGM reading where 75% of all readings fell at or below this value, after ranking all values from lowest to highest.
  • IQR: Inter Quartile Range (IQR) is the difference between the Quartile 75 and Quartile 25 readings.
  • Std. Dev.: Shows how much CGM glucose readings rise and fall. This is also known as glycemic variability.
  • IQ Std. Dev.: Inter Quartile Standard Deviation (IQSD) removes the top 25% and bottom 25% of CGM before calculating the SD.
  • SD Mean: Standard Deviation of the Mean (SD Mean) estimates variability by dividing the SD by the square root of the number of values.
  • %CV: The Coefficient of Variation (%CV) is calculated by dividing the glucose Standard Deviation by the mean glucose. %CV is a standardized measure that assesses the magnitude of glucose variability. The larger the %CV, the larger the variability in CGM readings.

Standard deviation
In statistics, the standard deviation is a measure of the amount of variation of a random variable expected about its mean.[1] A low standard deviation indicates that the values tend to be close to the mean (also called the expected value) of the set, while a high standard deviation indicates that the values are spread out over a wider range. The standard deviation is commonly used in the determination of what constitutes an outlier and what does not.

I searched, I’m guessing like you did, and didn’t find any recommended ranges for the IQR statistics. I did find a paper that has a very readable discussion of how IQR and IQR SD is useful. Optimizing Display, Analysis, Interpretation and Utility of Self-Monitoring of Blood Glucose (SMBG) Data for Management of Patients with Diabetes

I did a couple of comparisons of good weeks versus not good weeks in my own Clarity data. The IQR data wasn’t telling me much I didn’t already know. I wonder if it is more informative when TIR is low.

As we can see from the Clarity doc SD mean isn’t average SD. 1-2 means very little variability. Similar to SD by itself the value isn’t useful because you can be out of range with very low variability and its still bad. Combined with the min, max and mean it numerically tells you the story that the the graph is visually communicating.

I’ll close this out with some statistics fun. Turns out the moral of Aesop’s story about the boy and the swallow is never never look to statistical outliers, especially on a gaussian distribution, in order to understand the mean. Credit to Adam Rutherford in Natalie Haynes Stands Up For the Classics 05 Aug 2024 23:50-25:00

3 Likes
3

@spdif
Thank you so much for a thourough answer.

I´ve been looking for the definitions from Dexcom but even Dexcom support couldn´t provide them. Strange, but problem solved so thanks again.

I´m looking forward to reading the paper “Optimizing Display…”.

Knowing all the definitions I can finally find ways to use the numbers. I will use the IQR, the median and the quartile 25 and 75 numbers. Especially when I know my Dexcom have had many false lows and highs. This happens from time to time, so IQR will come in handy.

Personally I think I´ve found my limits for how tight control I can have with the tools I have avilable.

I try to get my median value to max 125 pr hour and the SD to max 35.
I aim for 90% TIR and are mostly getting 95% or more.

I got good advice at this forum many years ago and that was:
-use 14 days of data at the time
-look at the average or mean BG
-SD should be half of your average BG value
-identify only one problem at a time
-fix the lows first
-use the visuals (AGP or Dexcom trends) to identify your biggest problem
-set one goal
-repeat after 14 days

Everyone is more than welcome to chime in with anything useful related to using data to our own benefit.

1 Like
4

You are welcome and I’m glad the information helped. I’ll be the first to point out I didn’t answer most of your questions so I hope others will share how they are evaluating the metrics Clarity offers.

Let me know if you need more info on SD Mean. I can ask some math professors I know to help us make sense of what it represents.

5

I agree with this. The other statisitic are interesting, but, in my opinion, focusing on the AGP numbers are the most beneficial. If your numbers on the AGP are good, likley the other statistics will be fine.

Some users may use the other statistics to fine tune control, but for vast majority of users focusing on the AGP would be good.

1 Like