In one sentence
The standard deviation divided by the mean, as a percentage: the variability figure that does not depend on how high your average is, and the one VT uses to compare months.
Why it matters
It answers the question the SD cannot: how much does my glucose swing, relative to where it sits? In diabetes care it is the agreed measure of stability. In 376 people with diabetes studied with sensors in Montpellier, the upper limit of the CV in the most stable group, people with type 2 diabetes on diet or insulin sensitizers alone, was 36 percent, and above that line low readings became markedly more frequent in insulin-treated people; the authors proposed 36 percent as the threshold between stable and unstable glucose (Monnier 2017), and the international consensus adopted it (Battelino 2019). Healthy people sit far below it: a mean CV of 17 percent, give or take 3, in the 153-person reference study (Shah 2019). For a reader without diabetes the useful comparison is with your own months, where a rise of a few points usually has a visible cause in the meal log.
What moves it
Upward: big or late meals, erratic timing, poor sleep, an evening's alcohol, being unwell, hard training on an empty stomach; and the sensor's own noise on a day with low coverage. Downward: regular meals of moderate size, protein and fiber with carbohydrate, a walk after eating, enough sleep.
How to read it well
Over a period, not a day: the month's CV is the figure, and the day's is a detail. Compare months of similar coverage on the same sensor site.
How Vitality Tracker uses it
Calculated per day and per period and shown with both reference points named for what they are: the diabetes-care threshold of 36 percent and the healthy-adult mean of 17 percent. The Baseline Month's CV is the reference for later months. It is not a Vitality Score input.
How far to trust the number
Calculated from measured readings; robust for a period with good coverage, inflated by artifacts on a day with few readings.
Sources
- Monnier L, et al. Toward defining the threshold between low and high glucose variability in diabetes. Diabetes Care 2017;40:832–8. doi:10.2337/dc16-1769
- Battelino T, et al. Diabetes Care 2019;42:1593–603. doi:10.2337/dci19-0028
- Shah VN, et al. J Clin Endocrinol Metab 2019;104:4356–64. doi:10.1210/jc.2018-02763.